cyphersolver

An informal benchmark of frontier AI models on historically "unsolved" ciphers, as much a hobby as a measurement. Most of the work is by Claude (Opus 5 and 5.5, some Fable 5.1); a few results came from outside contributors using other models. Each target's profile.json records which model did what, and what it was given. The first targets came from three standard lists: S. Tomokiyo's Unsolved Historical Ciphers, Klaus Schmeh's Top 50 unsolved encrypted messages, and Elonka Dunin's famous unsolved codes. Since mid-September 2026 most come from this project's own catalogue of undeciphered archival letters (DECODE records, BnF and Gallica notices, Tomokiyo's "undeciphered" letters, printed calendars). Each one is a test of how far an AI assistant, working with a human, can get on a problem that has resisted people for decades or centuries: how much is archival research, how much is cryptanalysis, and where it stops. Method: archival research, historical cribs, printed editions, and small purpose-built solvers, always run against matched controls so that a negative result says something. The results are being written up with George Lasry as a paper on how LLMs perform against historical ciphers.

• Contributing: the ledger takes contributions from anyone: a solution to a listed cipher, a new unsolved cipher, a new cipher with its solution, or an independent check of a reading. See CONTRIBUTING.md and the site's How to contribute; everyone credited is on Contributors, built from outcome.credits in each profile.json and the registry contributors.json.
• Website: https://dbourdeau.github.io/cyphersolver/ — hub, priority queue, and formal write-ups (source in docs/).
• Solved, catalogued: SOLVED_CATALOGUE.md — every item solved, read or partly read, the method that broke it, and where the AI did the work.
• Solved, ranked: SOLVED_RANKING.md — the finished targets in one table, scored by difficulty, historical weight, novelty, research effort, profile and verification.
• Catalogue of open targets: CATALOGUE.md · site page — undeciphered letters and series absent from the standard lists, scored for importance, solvability and difficulty. Generated from catalogue.json; a target leaves it when it is read, resolved or found in print. How it was compiled and refilled (BnF/Gallica, DECODE, the 22 Sept 2026 new-solves sweep) is in CATALOGUE.md; harvest scripts in research/catalogue_harvest/.
• Per-target record: each folder under targets/ has a NOTES.md (sources, dead ends, what is established and what is inferred) and a profile.json (the fixed-field record behind the paper; profile.schema.json).

At a glance

277 rows record a reading of some kind; the site has 318 write-up pages and 60 catalogue targets are open. Counts are rows of the tables below (a row can cover a group of letters), regenerated by docs/_build_site.py.

Outcome	Rows	Complete
Key recovered from ciphertext-only	47	26
Key recovered based on plaintext from external sources	13	9
Key recovered based on adjacent plaintext	73	36
Read after matching with key from external sources	21	14
Read with known key	83	52
Read from existing decipherment	40	33
Not solved	49	—
Not applicable	18	—
In progress	11	—

Results

The third column of each table is the day the finding landed in this repository: the date of the first commit that carried it, taken from the git history, not the date of the document. The write-up pages carry the same date, and docs/_dates.json holds them all. "Date" is the date of the document itself.

Key recovered from ciphertext-only

Key rebuilt here from the ciphertext alone. Independent re-solutions of items solved elsewhere are marked.

Target	Date	Landed	Extent	Result	Where
Unknown sender → unknown recipient, undated Latin diplomatic letter, Riksarkivet, Stockholm, Chifferklaver låda II:113 (DECODE R4282; separated from catalogue 207, class A)	Undated; catalogue 1600–1699	14 Oct 2026	complete	Contributed by George Lasry. Transcribed with CTTS (CrypTool Transcriber and Solver); key recovered from ciphertext-only with the CTTS built-in homophonic solver. GPT-6 Astra improved and finalized the key and decryption using the images and four cleartext spans; no historical decipherment or external key was used. Both pages are deciphered with three unidentified name/title signs (five occurrences); the key is partial. Key-assignment coverage: 800/805 non-null signs (99.38%), or 1,079/1,084 including 279 nulls (99.54%); these are not coherent-reading percentages. The earlier grouped attempt remains as historical context.	targets/decode4282/ · write-up
Giovanni, cardinal Salviati to Not established, 16 October 1525 (Français 2933)	16 October 1525	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/8 pages transcribed; 1 more in part; 3/8 pages deciphered in part. 838 elements, 704 identified; 2 further groups with unresolved boundaries; transcription. Spot-checked here 13 Oct 2026: partly confirmed.	targets/salviati1525/ · write-up
Marguerite Paléologue (catalogue attribution for ff1 onward) to Louis de Gonzague, duke of Nevers (catalogue attribution), 1562–1564 (catalogue group; individual dates unchecked) (Français 4687 — ff5v–7v)	1562–1564 (catalogue group; individual dates unchecked)	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 4/4 pages transcribed in part; 2/4 pages deciphered in part. f5v/f6r: 241 elements transcribed and deciphered. f6v/f7v: 356 base signs transcribed, none. Spot-checked here 13 Oct 2026: partly confirmed.	targets/paleologue1562/ · write-up
Gasparo Fogliani to Not established in the checked postscript, 20 October 1568 (Français 3234)	20 October 1568	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/2 pages transcribed; 2/2 pages deciphered. f103r–v: 422 elements, 418 identified (99.1%). Spot-checked here 13 Oct 2026: confirmed.	targets/fogliani1568/ · write-up
René de Lucinge to Charles-Emmanuel I, Duke of Savoy, 9 June, 18 June and 3 August 1588 (Italien 2242 — Lucinge, 1588)	9 June, 18 June and 3 August 1588	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 7/30 pages transcribed; 1 more in part; 7/30 pages deciphered; 1 more in part. Transcribed passages: 14,608 elements, 14,306 identified. Spot-checked here 13 Oct 2026: partly confirmed.	targets/lucinge1588/ · write-up
Méry de Vic to Henri IV, 28 August 1589 (Dupuy 61 — Méry de Vic, f43r)	28 August 1589	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 6/6 pages transcribed; Partly deciphered. 545 elements, 523 identified (96.0%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/meryvic1589/ · write-up
Juan Blasco de Aragon to Not established, 2 May 1597 (Français 23054 — Blasco, 2 May 1597)	2 May 1597	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 3/4 pages transcribed; 1 more in part; 3/3 cipher pages deciphered. 107 elements, 102 identified (95.3%). Spot-checked here 13 Oct 2026: confirmed.	targets/blasco1597/ · write-up
Louis-Sauveur, marquis de Villeneuve to Jean Otter, 29 June and 19 August 1737 (Français 22225)	29 June and 19 August 1737	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 6/6 pages transcribed; 5/5 cipher pages deciphered. 815 elements, 811 identified (99.5%). Spot-checked here 13 Oct 2026: confirmed.	targets/villeneuve1737/ · write-up
Writer in Mondesir, Isle de France (name not established) to The writer’s wife, 20 November 1769, with later additions (Français 5581 — Letter from Mondesir)	20 November 1769, with later additions	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 13/13 pages transcribed; 11/11 cipher pages deciphered. 4,923 elements: 3,950 deciphered, 949 separator marks (?), 24 unidentified. Spot-checked here 13 Oct 2026: partly confirmed.	targets/mondesir1769/ · write-up
Napoleon III / Empress Eugénie (both directions) to Empress Eugénie / Napoleon III (both directions), June 1859 (NAF 11355 — Emperor–Empress code)	June 1859	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 15/15 pages transcribed; 9/11 cipher pages deciphered; 2 more in part. 440 elements, 370 identified (84.1%); 4 assumed readings; 50 ?. Spot-checked here 13 Oct 2026: partly confirmed.	targets/eugenie1859/ · write-up
Jean, cardinal Du Bellay to The Constable, Undated (Français 3142 — Du Bellay, ff65–66)	Undated	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 4/4 pages transcribed; 1/4 pages deciphered; 3 more in part. 851 elements,547 identified (64.3%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/dubellay65/ · write-up
Not established to Not established, 18th century? (Français 14765 — p351)	18th century?	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: Transcribed; Deciphered. 936 elements, 873 identified (93.3%); 61 ?, 2 unidentified. Spot-checked here 13 Oct 2026: confirmed.	targets/fr14765p351/ · write-up
Unidentified to Unidentified, Undated (Français 3575 — f56r)	Undated	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: Transcribed; Deciphered. 782 elements, 774 identified (99.0%). Spot-checked here 13 Oct 2026: confirmed.	targets/fr3575f56/ · write-up
Desiderio l’Abbé → Duke of Nevers, Prague and Breslau, BnF fr. 3198 nos. 30, 37; second copy no. 38 (catalogue #284, class C)	2 Mar and 2 May 1577	7 Oct 2026	complete	Key recovered by AngusRobinson (PR #20) from the ciphertext with the probable word l’empereur: homophonic figures, vowels 1–15, consonants two-figure groups, 6 = word space. All 31 runs deciphered, 534 of 536 tokens (99.6%); checked here against the images, the 23 Sept draft (98.1% of figures agree) and 1,000 shuffled keys (z = 12). Polish court politics in March; in May, of Weber: “mais plusieurs ne pleureroient s’il n’y retournoit jamais”. Two blotted figures restored from context.	targets/labbe1582/ · write-up
George W. Erving → James Madison, Madrid, private No. 24 "Duplicate", "In Mr Pinkneys Cypher, the Cypher of ye Legation", NARA RG 59 Despatches from US Ministers to Spain vol. 10 = M31 reel 12 frames 0362–0367 (NAID 188605361; Founders Madison/99-01-02-1993, early access; not in the catalogue or on DECODE; sibling of targets/erving1807/)	10 Aug 1807	4 Oct 2026	complete	Key recovered from the Founders text and the code of the NARA duplicate; contributed by Feyseel Nur (PR #21), with Claude Opus 5.5 subagents; adversarial review by OpenAI Codex. Founders prints the letter decoded, with ten blank gaps and without the groups; the duplicate on the film has the groups and no decode. 1,145 groups transcribed and aligned with the Founders text: Pinckney's legation code rebuilt to 414 rows, partly alphabetical (870–880 their … thing, 1340/1341/1343 an/ance/and). Checked against Jacob Wagner's 1803 decode of Pinckney's own despatch (M31 reel 7 frame 0348): 45 of 47 shared groups agree, 584 and 943 undecided. 1,125 of 1,145 groups (98.3%) give sense without the 11 conjectures (1,136 with them); Codex: confirmed but overstated, 716 groups (62.5%) have outside support (Wagner, the March letter), 908 (79.3%) recurrence or outside support, 0 digit errors in 53 sampled groups, and a conservative count without hidden-digit restorations and emendations of 1,081/1,147 (94.25%). Six of Founders' gaps filled from the code (that government, and, hostile, resort, to Hanover and, these), expectations plausible, past and pride guesses; three disagreements (very [?] authority for "best informed persons", the vast [3 groups] Spain by the war for "costs brought about", no with). Godoy paid for the Hanover bargain "out of his own funds", is "deceived and furious", and "as a dernier resort he will make an alliance of some sort with England!". Open: 633; 632, 455, 1482; 767, 679, 378; one group in the binding; 916 in the postscript; two uncertain positions on frame 0367.	targets/erving1807aug/ · write-up
Archduke Leopold Wilhelm → the abbé (baron) de Mercy, instruction in cipher from the camp at Barneton, BnF Espagnol 144 f. 22r–v (Gallica btv1b10035717h; finding-aid item 11; not in the catalogue, GitHub issue 16 by NoAutopilot)	6 Jun 1648	1 Oct 2026	partial	Key recovered from the ciphertext alone by NoAutopilot (cipher-lab); transcription checked on the images and corrected here. A homophonic substitution of the numbers 2–34 inside clear Spanish; 521 of 529 cipher tokens read (98.5%, measured). The four "uncertain codes" 48, 52, 65, 72 are digit pairs (4 8, 5 2, 6 5, 7 2) and two 26 are 2 6; the square-with-a-dot sign is the comte de Saint-Ibal (the clear instruction of 13 Apr on f. 21r asks that Santibal come with Mercy); "su camarero mayor" is Konrad von Burgsdorff, Brandenburg's chief chamberlain. Mercy is to go to Cleves and ask the Elector whether 3,000 infantry in two or three regiments can be raised (the substance of Lonchay et al. IV no. 183). Open: eight tokens of f. 22v line 4 (code 15) and the letters lost at the trimmed edge of f. 22v	targets/mercy1648/ · write-up
Otto von der Malsburg, general war commissary → Landgrave Wilhelm V of Hesse-Kassel, ten cipher letters on the war in Westphalia, Hessisches Staatsarchiv Marburg, HStAM 4 h Nr. 1411 ff. 3–33 (HCPortal 496–509, all "Not solved"; catalogue 338, class C)	15 Jan – 28 Mar 1637	28 Sept 2026	partial	Key recovered from the ciphertext alone; the letters deciphered. No Malsburg key among the 176 HCPortal key records of the Marburg key volumes (harvested). German homophonic table of 89 two-digit groups (2–7 homophones per letter) annealed from three letters, all restarts converging; digraph signs (N ch, Z st, W ss, L tt ...) valued from context. 9,338 of 9,736 cipher groups (95.9%) decipher to sense, 97.9% of the non-code text (shuffled-key control 50–58%); 236 doubtful groups re-checked on the scans. Garrisons without pay, Ehrenbreitstein, safe-conducts refused for the Cologne peace talks, the Dutch alliance. Also a French letter signed "232" and London news extracts in the same key. The separate ten-line alphabetic block on f. 12 now has a recovered period-five key, contributed by Larry Beck with ChatGPT: shifts 4, 5, 3, 6, 2 over 24 letters. Its selected transcription contains 510 alphabetic positions; its critical reading is partial and is not included in the numerical 95.9% measure. Open: about 40 numerical code groups, alphabetic groups 43, 45 and the angular sign labelled LL, and local manuscript readings.	targets/malsburg1637/ · write-up
Sienese envoys and others → the government of Siena, the cipher letters of the Concistoro, Archivio di Stato di Siena, Concistoro 2308 fasc. 2 "Lettere in cifra" (DECODE R4790–R4814, all dated "1500–1599", no status; keys R4746–R4789; catalogue 336, class B)	before 1429 to 1547	24 Sept 2026	partial	Keys recovered from the ciphertext alone for three letters, and from a fragment on the letter for a fourth. None of the 43 filed keys fits any letter. No. 25 (a lord's route by Lajatico, Radicondoli and Casole into Sienese territory): 477 of 496 signs (96.2%). No. 18 (Latin treaty articles with the Emperor, subscribed Niklas Ziegler, war on the Florentines at Ponte a Valiano): 1,271 of 1,459 signs (87.1%). No. 14 (envoys writing from Rome while Carlo Malatesta, d. 1429, was alive): 58.6% in sense, 29.2% doubtful. No. 4 (Galgano Borghesi to Leonardo Benvoglienti, Naples, 12 May 1456): 114 of 144 signs, key from an alphabet fragment below the text and the recipient's glosses. Nos. 8, 10, 13/16 and 22 deciphered at the time. Open: Meister's 1421 letter (no. 1) and nos. 2, 26–28, not photographed; two long systems (nos. 6 + 24, 20 + 23); nos. 7, 9, 11, 15, 17, 19, 21 without a key	targets/siena1421/ · write-up
Cristofano → the Signori of Siena, two letters on the war over Lucca, Archivio di Stato di Siena, Concistoro 2308 fasc. 2 nos. 6 and 24 (DECODE R4795, R4812, both catalogued "1500–1599"; key R4755 in fasc. 1; split from catalogue 336)	c. 1431–1438 (no. 6: 19 February)	14 Oct 2026	partial	Deciphered by George Lasry (with Claude and the help of ChatGPT Astra), sent to Daniel Bourdeau 14 Oct 2026: key recovered from the ciphertext alone (homophonic, 36 signs for 21 letters, 7 nulls, 18 word signs), then matched to the filed key R4755. Verified here on our blind transcription of 24 Sept: same letters at 69% with the key read by eye, 88–91% after fitting our labels; LM −3.06/−3.34 against shuffled-key −4.03/−4.04. No. 6 95.4% (after the z/u key correction, 14 Oct) and no. 24 79.6% (revised 14 Oct from 72.7%) of decrypted letters understood; the dorse of no. 24 decoded here in part and signed "Crist[ofa]no" (same writer). Piccinino opens the road to Lucca; Mantua and Taliano go over to Milan. Dated c. 1431–38, not 1431–32 (Mantua marquisate Sept 1433)	targets/siena1431/ · write-up
Matheo de Segura → Constable of Castile, Madrid, BnF espagnol 336 no. 99 ff. 196r–198r (catalogue 321, class B)	8 Aug 1596	23 Sept 2026	complete	Both short passages; all 125 cipher tokens proposed, 119/125 (95.2%) excluding six tentative assignments. Homophonic key rebuilt by annealing and image resegmentation. The world ending or the king dying by September, and Ponte/Aponte with Escudero. No independent key; punctuation and name segmentation editorial. Previously the last item listed undeciphered in Tomokiyo’s es. 336 list after Lasry’s 2023 solutions (ff. 69/81 printed keys, f. 101 interlinear decipherment, f. 148 near-monoalphabetic); Devos 1950 full tables not obtained.	targets/segura1596/ · write-up
Giovan Gioacchino da Passano (attributed), London → Francis I, BnF Clairambault 331 ff. 156–157 (catalogue 281, class C)	15 Mar 1530 (copy heading)	23 Sept 2026	complete	Surviving account deciphered. Dotted re-transcription overturned the failed attempt: cash, credits, annual revenues, jewels, gold/silver vessels, furnishings, Winchester, St Albans and York. 395/402 signs assigned; 386/402 coherent (96.0%) after 13 explicit emendations. Three rare signs remain open; a page and three lines omitted by the historical copyist are absent and unread, outside the denominator. Partial key; no independent plaintext.	targets/joachim1530/ · write-up
Unknown Catholic gentleman of the Agenais → the connétable de Montmorency, "Mémoire en chiffre du XIIe decembre 1560", BnF fr. 3157 no. 67, ff. 152r–156v (catalogue gives f. 158; Gallica btv1b90598645 views 153–158; catalogue 276, class C)	12 Dec 1560	22 Sept 2026	complete	Lasry solved it independently in 2023 (improved Feb 2026), unpublished (private communication). Solved: key broken here ciphertext-only; read (98.9% of 1,715 cipher words, measured). Ten pages wholly in a cursive pseudo-script, no decipherment or key anywhere. Homophonic substitution with word division, two-stroke units (zo = u, pp = m, s3 = r, lz = d), polyphonic δ (u/b/g) and four word codes (Z = pour, 4 = roy, c+ = et, F = sieur); the transcription corrected against the images with the key (about 300 x→r fixes). A report on the Protestant rising in the Agenais a week after François II's death: 2,000 at Clairac, the Caumonts and their ministers, Biron turning back from Bergerac, the rebel captains at Nérac around Navarre ("le chief"), the Agenais deputies to the États of Orléans. Open: 16 scattered words; the writer's name (not in the text).	targets/memoire1560/ · write-up
Unknown author → unknown recipient, "L'Estat du Roy de Navarre et de son party en France", after 1580, BL Harley MS 1582 ff. 263–264 (DECODE R8505, "unknown sender", Non-decrypted, language "English?"; catalogue 121, class C)	after 1580	22 Sept 2026	complete	Cipher broken here from the ciphertext alone; read (95.1% measured). The catalogue item's other record, R8501, is a Stafford 1586 despatch already read from its own f. 72r decipherment (targets/harley1582r8500/). R8505 is this French memoir on Navarre's party, identified here from its clear text. Homophonic alphabet of ~50 graphic signs: the letter m is a group of two, three or four dots, taken for punctuation by every earlier pass, which is why no key fitted. A control experiment (real French under a key of the same shape) located the fault in the solver, not the manuscript: with a wrong Sukhotin constraint and the temperature scale fixed, the control solves 100%, and the real text broke at once. Reads: the prince of Parme "pour lui rendre les places", "un party de malcontens", an "italien qui sert d'ingenieur a Lisbone", "une armee de reistres", Beaucaire sur le Rosne, Béziers, Carcassonne, Montmorency, d'Épernon. Open: ten nomenclator numbers (names; 72 is a country, almost certainly France, 81 a person, probably the King)	targets/harley1582r8505/ · write-up
Ferenc Wesselényi, Palatine of Hungary → the King (Leopold I, inferred from "Maiestatis Vestrae"), Latin draft with cipher runs, MNL OL E 199 (Wesselényi family archive), 8. pallium 1. 01 (DECODE R674, "non-decrypted", dated 1 Jan 1664 = start of a 1664–1670 span; catalogue item 45, class C)	c. 1663–64	22 Sept 2026	complete	(241 of 244 cipher tokens, 98.8%, measured). Broken ciphertext-only: a Latin anneal allowing two numbers per letter, then hand correction from the clear context. The cipher is a reversed alphabet, 24 = a … 3 = x, with vowel homophones 25 a, 26 e, 28 o, 29 u; DECODE key R672 (Teleki collection keys no. 12) has the same alphabet, found afterwards, but its syllables and codes do not fit. The general [60] should join the regiments of Heister, Schneidau and Sporck, add 2,000 foot and camp at [64] so that it takes His Majesty's garrison, by force if need be: "ista nux durissima … frangi debet". Struck corrections: Wesselényi's draft. Open: codes 60 (the general, perhaps de Souches) and 64 (the place), no key	targets/wesselenyi1664/ · write-up
Sir Simonds D'Ewes, private cipher log of his son Clopton's convulsion fits, BL Harley MS 286 f. 61 (DECODE R7759, "unknown sender to unknown recipient, 1635"; BL catalogue: Harley MS 289 f. 61r; catalogue item 130, class B)	3 Apr 1635 – Jan 1636	22 Sept 2026	complete	(96.5% of 287 words, measured). D'Ewes's own invented alphabet, simple substitution of Latin; key rebuilt by crib (convulsio, plena luna, difficilis), weekdays and moon phases checked, Autobiography confirms 3 Apr 1635 start and 44 fits. Open: last line trimmed under the mount, one word of l. 13, a faint phrase in ll. 24–25.	targets/harley286/ · write-up
Unknown writer → Hesse-Kassel, Hessisches Staatsarchiv Marburg, HStAM 4 d Nr. 1218 no. 49 (DECODE R4500, "partially decrypted"; catalogue 153, class B)	undated (volume 1635–52)	22 Sept 2026	complete	Digit-and-sign letter substitution (12 = r, θ4 = ch, a few homophones), no sibling key in the volume's ~120 key sheets; annealed on de-1500s, confirmed by the contemporary gloss on lines 1–4. ~97 % read (report to the Emperor and the Elector of Bavaria, "damit ich aus allen fall entschuldigt sein"); nine signs at the end of line 1 open, under the faded gloss	targets/marburg1635/ · write-up
Ferdinand of Bavaria, Elector of Cologne (Bonn) → Duke Maximilian I of Bavaria, ten ciphered pages, BayHStA, Kurbayern Äußeres Archiv 4591 ff. 281–286 (DECODE R9426, "non-decrypted", no author, no recipient, no place, no language) (catalogue 165, class C)	22 Dec 1619	22 Sept 2026	complete	(95.8% of cipher tokens measured). Broken ciphertext-only: a homophonic substitution on the even two-digit figures 10–78 (vowels with two or three homophones, consonants one, k and w only in the upper series), two signs for word division, and a three-digit nomenclator alongside; key recovered by annealing 2,482 letters against lang/de-1500s, with ~60 single-digit transcription slips repaired (0/6, 3/5, 1/4, 2/7) and a sample re-checked on the images. Control: the key reads the sibling R9425 unchanged ("Albrecht von Gottes Gnaden"); on 2 Oct 2026 R9425 (ff. 278–280) was transcribed and deciphered in full with the key unchanged, 97.3% of 1,567 cipher tokens measured: Archduke Albert's written resolution of 15 December 1619 for the League's agent Licentiate Morreus (Bohemia to be recovered, his troops to join the Catholic army, no muster place or 500 horse because his levies go to Luxembourg, Breisach referred to Oñate, the 1,000 cuirassiers to Ambrogio Spinola); it shows 414, 415, 416, 417, 419 to be nulls and fixes 242 = und, 165 = Mt., 2 = Ambrosio. The record's "unknown" sender is named by the clear last page — Bonn, 22 December 1619, signed Ferdinand — and by the ciphered "mein Thumbcapitel", "meine Landtstendt", Paderborn. Six weeks after Frederick V's coronation: quartering companies in the county of Horn, at Montfort and at Deutz, two Kopfstück or two Ort of a Königsthaler a day, Lintelo's and Herberstorff's levies, a year's respite for Paderborn, and the argument that unless the Catholics act they will shortly have nothing left to secure them against "Einfal, Gwalt und Betrangnus". Not in Briefe und Akten NF I/1 or any other edition searched. Open: 31 of the 32 nomenclator groups (only 248 = und established; no key for this cipher survives in the volume), three unidentified names, four words	targets/ferdinand1619/ · write-up
Charles of Egmond → grand master of France, Arnhem, BnF fr. 3015 no. 8, fol. 16 (catalogue 28, class C)	18 July (?), year unknown	20 Sept 2026	complete	Independent re-solution, not a first break (Lasry 2023). Read, with minor uncertainties; already solved by Lasry in 2023 (key on Tomokiyo's GL.htm §2, pointed out by Lasry 21 Sept 2026), so this is an independent re-solution. Sixteen cipher lines, two insertions and the address recovered as French with an inferred substitution key and nulls. Requests support after entering war on the king’s commands and assistance for the commander of Saint John. Date sign and a few literal anomalies retained; no independent key or clear-copy control, no personal recipient name or year established	targets/egmond/ · write-up
Swatow telegram to Sun Yat-sen (JACAR B03050738800)	1916	15 Sept 2026	complete	Systematic code condenser over the standard telegraph code recovered by brute force; 41 of ~44 characters read	targets/sunyatsen/ · write-up
Sun Yat-sen's intercepted telegrams, JACAR 特定人ノ発受郵便物取調 (Mar 1916 – Mar 1917)	1916–17	18 Sept 2026	partial	The Japanese Ministry of Communications' copies of Sun's circle's wires: ~60 condenser telegrams read under seven keys of the Swatow family (Sun ↔ Yamada, Chen Qimei, Qingdao, Manila, San Francisco, Dai Jitao 1917), plus the Hankow +111 digit code; Chen Qimei's assassination as reported to Tokyo. The July–November 1916 telegrams are in two other systems, decoded by Brice Dudout (Oct 2026; checked here on three telegrams); Dudout also showed that most of our one-letter emendations were our own misreadings of the forms, not clerks' slips; the June 1917 telegrams are open	targets/sunyatsen/ · y1916/ · y1917/ · write-up
Warsaw, 24 Dec 1627, to the Bishop of Olmütz? (DECODE R1408, ÖStA HHStA)	1627	16 Sept 2026	complete	Read in full from Tomokiyo's transcription: a homophonic substitution whose alphabet turns out to be in plain order (odd figures a–m, even figures n–z), with letter pairs for consonants, a dozen two-figure syllables, the letters a m as nulls and thirteen word codes. A 5-gram + dictionary annealer converged from random and seeded starts, shuffled controls failing. The writer, back in Warsaw from Nikolsburg, presses for a promised canonry of Olmütz for one of the Queen of Poland's sons. 2 Oct 2026: the DECODE image checked (Tomokiyo's hapax 12 + null o is the code 120; a gap on line 10 is a 29, stimato; a docket at the foot names Card. Dietrichstein, the addressee), measured 316 of 332 significant tokens (95.2%); the sibling candidate R1409 (f. 178) is a different system (the R1408 key ranks 67th of 500 random keys on it, a ciphertext-only attempt did not converge), and none of the fascicle's key records R1392–R1406 is this key. Ten word codes and two small groups stay glossed from context only; no key material	targets/warsaw/ · write-up
Henry of Navarre → Ségur, BnF 500 de Colbert 401 ff. 233, 239, 288v (listed as Henry III)	1585–86	16 Sept 2026	complete	Lasry solved it independently in 2024, unpublished (private communication). Three figure-cipher letters to the envoy raising a German army, read from the Gallica images. The upper figures fall into blocks of five (mod 5 test), an alphabetical syllabary; a structured annealer with the letters' clear French as context recovered a key that is alphabetical on its own (a 4–6 … u 45–47, x y z, et, null; syllables ba–vu 54–123; doubles ée ll ss), 97 % on a matched control. Conclude with Duke Casimir, raise the largest levy and march it; Clervant's two thousand reiters; the Vivarais or Dauphiné road. Word-signs partly glossed; f. 143 (1583) is another key. The whole volume swept: one more cipher leaf, f. 366 (autograph, same key, mostly names), and f. 333 read with Tomokiyo's sibling key, which is this key shifted by eight	targets/segur/ · write-up
Cardinal Soglia → nuncio Viale Prelà at Innsbruck, ciphered dispatch printed in L'Italia del Popolo no. 40 (Milan, 30 June 1848; Roman broadsheet reprint; Cipherbrain 13 Dec 2014)	1848	18 Sept 2026	complete	(18 Sept 2026): 5 = word separator, a 64-cell two-digit table of letters, homophones and syllables, 8XXX = one-part code in alphabetical order; Joe's transcription with one digit restored. A counter-order: stay with the Emperor at Innsbruck and suspend the earlier order to ask for passports after Lützow's departure; to be shown to Mgr Morichini. The DBI has the nuncio leaving Innsbruck on 14 June, a day before. Ten code words read from context and checked by alphabetical rank; 11 code groups open (the date of the earlier letter among them)	targets/soglia1848/ · write-up
Gottfried van Swieten (Bonn) → Count Cobenzl (Brussels), three despatches with ciphered passages in numbers, 1757–14 Jun 1759, ARA Brussels, Secrétairerie d'État et de Guerre 1236 ff. 141, 327–328, 344–345 (DECODE R955–R957, "non-decrypted"; catalogue item 39, class C)	1757–59	22 Sept 2026	complete	(22 Sept 2026). R955's 89 groups are deciphered in clear on the same page (Hanau magistrates and neutrality) and show an alphabetical syllabary. The 1759 letters use a new table (100–442), rebuilt ciphertext-only: monotone annealing against a French LM (one seed in eight broke it), then hand correction from the clear text round each run; four DECODE runs re-read from the scans. 96% of 553 groups read: the Gueldre convention and Moers, Soubise's army living on Gueldre contributions, Broglie the only competent general, complaints against Count Pergen. Open: about 21 groups, mostly names (232, the prince 147 100 126 148, one marshal).	targets/swieten1757/ · write-up
Johann Heinrich Kauderbach (Saxon-Polish resident, The Hague) → Brühl / Friedrich August II, ten intercepted despatches with passages in figures, 23 Dec 1754 – 25 May 1756, KHA The Hague, Prins Willem V inv. 201 (DECODE R1043, R1044, R1954, R1958, R1959, R2078–R2082, "never solved"; catalogue item 216, class C)	1754–56	19 Sept 2026	complete	(19 Sept 2026). The figures run on unseparated: 5 and 8 are nulls, the rest two-figure codes (first figure 0–4, 6, 7, 9; second 1–4, 6, 7, 9), one per letter or syllable, one key for all ten. Homophonic solves failed; Kauderbach's 1761 key in Dresden (DECODE R936, HStAD 10024 Loc. 08236/11 Bl. 5) gave the vocabulary, and a one-to-one annealing search onto it converged twice to the same key, finished by reading (01 oi, 12 ai, 27 in, 66 pour). Seven letters transcribed from the images. All ten read at 65–97%: the Dutch augmentation debate of 1755, Yorke and the Saxon subsidy treaty, the Treaty of Versailles of May 1756. Open: the clear sigla (S./., Q./., P./. …), four rare codes, R1954 (weak transcription).	targets/kauderbach1754/ · write-up
Potocka and Mniszech → Jakub Dunin, Archiwum Sanguszków, DECODE R7524–7530 and R7534–7536 (catalogue 280, C; formerly N/A, placeholder dates on some records)	1714–16 and undated	23 Sept 2026	partial	Potocka alphabet broken from the Polish R7526 postscript and transferred to eight French letters, including doubled homophones. 863/887 Potocka tokens valued (97.3%, key coverage); measured as sense word by word (5 Oct 2026), 844/886 (95.3%) in Potocka's letters and 844/1,118 (75.5%) with Mniszech's R7524 unread. Property, court influence and correspondence security. Four person codes, figure 73 and doubtful spellings remain open; Mniszech R7524 is unread and solver failure is inconclusive. No prior reading found; existence unknown.	targets/potocka1714/ · write-up
Four cipher letters to Gianfrancesco Gonzaga, lord of Mantua: Pandolfo Malatesta, archbishop of Patras (Rome), an unnamed agent (Rome), an informant at Forlì and Francesco de' Cattanei (Rome); Archivio di Stato di Mantova, Archivio Gonzaga (DECODE R1854, catalogued as one letter from an unknown sender; catalogue 152, class B)	22 Apr 1428 – 21 Feb 1430	22 Sept 2026	partial	Read in part: three of four letters read. Rome 9 Aug 1428 (1,802 signs, homophonic + nulls, key rebuilt ciphertext-only, 96.6%): a Gonzaga marriage, "Carolus est nimis iuvenis ad talia". Forlì 6 Nov 1428 (894 signs, read at the time; key from the interlinear gloss, re-deciphered 97.2%): the Malatesta, Urbino and the pope. Cattanei, Rome 13 and 21 Feb 1430 (1,736 signs, key rebuilt ciphertext-only, 96.7%/100%): the archbishop of Patras offers his see to Venice, the Rimini envoys. Open: Pandolfo Malatesta's own three runs (262 signs), not a simple or homophonic substitution, no key (R1853 register keys do not fit; the 1438 cipher/translation pair in ASMn b. 1081/840 is not online); 92.2% of all signs read (4,330 of 4,694, 5 Oct 2026).	targets/r1854/ · write-up
Guido Rangone (Venice) → Anne de Montmorency, BnF fr. 3070 f. 81 no. 40 (DECODE R4249, catalogued as "Guido Ramgone"; catalogue item 181, class C) and the sibling fr. 3082 f. 42 no. 19 (not on DECODE); 18th-c. copies Clair. 331 f. 15, Clair. 330 f. 206	10 Jan 1530; 24 Dec 1529	22 Sept 2026	partial	(245 of 264 cipher signs, 92.8%, measured by measure.py, 5 Oct 2026; was a hand count of 87.9%). Homophonic substitution, about 50 signs for Italian (vowels 2–8 signs, one null), no key known; broken ciphertext-only by simulated annealing against it-cinquecento, then hand corrections. The sibling was found through the Clairambault headings ("Vol. 143 fol. 42" = fr. 3082). Both passages concern Cesare Fregoso, Rangone's new brother-in-law: "ha mandato a dirlo al signore Cesare e farli grande oferte" (Venice made him governor of Verona, Jan 1530) and "il Fregoso ... tolse fare la cosa de [?]en". Open: four short runs of signs that occur once (A1–A2, A6, B3, B4), blocked by length	targets/rangone1530/ · write-up
[Thomas Edmondes, at Henri IV's court] → Lord Treasurer Burghley, BL Stowe MS 166 ff. 50–61, 109–112 (DECODE R7770–R7775, catalogued as "Lord Threr? to unknown recipient"; catalogue item 140)	Jan 1593 – Jul 1594	21 Sept 2026	complete	Independent re-solution, not a first break (Butler 1913). DECODE's author field is the addressee. Six letters from the King's camp (Tours, Blois, before Laon), each with a few lines in invented signs. The signs are a simple substitution of English with homophones and person signs (9, n, £), broken from the clear context. Every cipher line reads, the four on f. 59 (R7772) since 2 Oct 2026 (126 signs, key unchanged but for a vertical stroke = q: "and by consequence popis[h]lie nourished, pretending so to appertaine to the place of the Constable, to haue the gouer[n]ment of the younge heire of the crowne"; its words all on Butler p. 56): the King breaking off his sister's marriage to [9], Montmorency's demand for the young Condé, "that his faith is not vendible", and [£] (Guise?) and Champagne and Reims in 1594. Butler, The Edmondes Papers (1913), prints five of the six letters (all but R7774) with the cipher deciphered; found 30 Sept 2026, after the key was recovered here. The four letters dated 1592 are 1593 new style (the year began on 25 March). 3 Oct 2026: R7774 and R7775 re-read on the images (f. 109v lines 1–2 added, "[9]: of whome especiallie [o] hath made choise …"; f. 112 "carryinge"); every passage transcribed letter for letter and measured: 1,023 of 1,035 signs read (98.8%), 99.9% without the person signs. Open: the person signs 9, n, £/£2, 4̇, o (11 occurrences; no key with names survives, Butler's page text blocked here) and one sign "3:" on f. 109v.	targets/stowe166/ · write-up
Claude de La Guiche, bishop of Mirepoix (Rome) → Montmorency, BnF fr. 3138 no. 22 f. 60v; with Noailles (Venice) and Seure (Lisbon), fr. 3151 nos. 33, 39–44 (catalogue item 10, class C)	22 Nov 1551 (1558)	21 Sept 2026	partial	La Guiche's 406-sign passage, a simple substitution of 22 invented signs, broken ciphertext-only by permutation annealing: Dom Diego and the Sienese envoy, the Count of Santa Fiora 'mal satisfaict' and ready to serve the king on conditions. One name and the struck-through opening open. Noailles 1558 read at the time (margin decipherment, cut by the binding). Seure's six Lisbon letters (~2,000 signs, homophonic) left open.	targets/guiche1551/ · write-up
An Orange prince (son of William V) → an officer of the Dutch émigré troops, Germany, KHA The Hague, Prins Willem V inv. 339 (DECODE R1892, "non-decrypted", "unkmown to William V, 1 Jan 1795"; catalogue 220, class A); the writer is Prince Frederick, in England	late Nov – early Dec 1795	21 Sept 2026	complete	Lasry solved it independently in 2021, unpublished (private communication). A 6×6 digit square broken ciphertext-only (21 Sept 2026). Each letter is a vertical pair of digits 1–6; quadgram annealing on the Dutch postscript gave the square, which read the French page unchanged (1,829 pairs). General Dundas, pay, embarkation and entry into British service of the rassemblement; "communiquer le contenu de ce chiffre à mon frère". 2 Oct 2026: R2242's word-sign values tried at every occurrence (zoo, als, dat, brief/lettre, zijn adopted; six ruled out) and f = aan found: 1,757/1,951 tokens read as sense (90.1%). Colenbrander, Gedenkstukken II nos. 724, 733, 734 (not printing the letter) give the writer, Prince Frederick (Nienburg, his command; in England by 26 Nov 1795), and the date, before Henry Dundas's orders of 8 Dec 1795 — not DECODE's 1 Jan 1795. 3 Oct 2026: six lines re-transcribed on the full-resolution images, 11 one-digit writer's slips emended, P = voor; stricter measure (context-only sign values count unread; y scored as ij; period-Dutch word check; French closing scored in French; shuffled controls 8.7/12.0%; the 2 Oct readings score 93.2% under it): 1,860/1,954 tokens read as sense (95.2%), p1 99.1%, p2 90.7%. Read (5 Oct 2026) under the project rule that 95%+ read as sense is a full reading; the Dutch postscript (90.7% on its own) keeps open recurring word signs (⊣H ×9, hook ×7, slash ×5 …); singletons have no key material.	targets/r1892/ · write-up
Charles III, duke of Lorraine → the comte de Vaudémont, Nancy, 18 Jun 1592 (BnF fr. 3621 no. 97)	1592	22 Sept 2026	partial	Independent re-solution, not a first break (Tomokiyo 2025). Partial, 76.6% measured (844 of 1,102 tokens, 5 Oct 2026; 71.5% on 22 Sept). DECODE R9449 Non-decrypted; no key published, but Tomokiyo's unsolved list (before 18 Sept) reports he read it in 2025 and quotes the opening, which our reading matches (found 24 Sept, not used); no. 22 is a plaintext decipherment with no cipher, and four fr. 3995 key sheets are eliminated. Broken ciphertext-only on 18 Sept (contributed, PR #4), then re-read line by line from the page with the key in hand: the system is a reciprocal letter-pair substitution in ordinary cursive letterforms (a/i b/p c/f d/q e/t g/u l/y m/z n/s o/r h/x), two dots for a doubled letter, a null word-divider, a few special signs, and syllable/word figures (31 = ma, 139 = vous, 145 = me). The earlier 44-symbol key hid this because the transcription merged e with c and b with h. 790 of 1,105 cipher tokens read (the 18 Sept key gave the same words for 40.6%): j'ay esté infiniment mal satisfait du peu d'execution … de tant de belles promesses … pionniers et munitions … le juste mescontentement … affin que 141 ne me delaissent tout à coup … pour y commander et conserver la plaine … de ramener mon armée ès quartiers de la Faulche … vous tiendrés les chemins … ma resolution, le peu de moien. Open: about 60 tokens in unkeyed special signs, ten single figures, three garbled runs	targets/lorraine1592/ · write-up
Unknown writer → "Monsieur", Nancy, December (17th c.; signed "Nicle"?), TNA SP 106/10 ff. 241–243 (DECODE R927; catalogue 56, class B), "undeciphered letter in the masonic alphabet"	17th c.	19 Sept 2026	complete	Independent re-solution, not a first break (Sheila Richards, Secret Writings, 1973, per Lasry). The pencil pigpen key on the flap (akt·blu·cmw/…) reads nothing in any orientation. The cipher is a homophonic pigpen (cell shape + 1–3 dots, inner-stroke variants; several signs for e, i, s, t, r). Hand transcription of the last page (450 signs) → quadgram solver finds French → key completed from words. Last page reads nearly throughout: je ne pouvois recevoir d'un ennuy plus sensible, ayant appris que du mariage de ce gentilhomme … l'entiere affection que j'ay vouée a vostre service … et vous supplie tres humblement, monsieur, me continuer l'honneur de vostre bienveuillance et croire que je vous honnore. 2 Oct 2026: ff. 241 and 242 hand-transcribed and decoded. The letter runs f. 243 → f. 241; f. 242 is a second copy of f. 243, line for line. f. 241 closes it: an envoy to congratulate Sa Majesté on recovering from a long illness, le zele que j'ay a son tres humble service, signed "Nicle"(?), De Nancy le [..] decembre. Measured: 84.0% of the letter's signs read as sense (f. 243 81.7%, f. 241 86.5%; the copy 73.5%); key extended (H3 = z, C2 = q, C1 f/q and F2 l/h polyphones). Neighbouring records R922–R933 opened: none in this pigpen. 3 Oct 2026: every sign of every unread word cut out at 3–4x and re-read (corner sides, dot counts, inner strokes, two writer-cancelled signs); F1 = l, D*1 = o; a joint re-anneal over-fitted and was not adopted. f. 243 now reads lequel salua dernierement [name] de la part de [name] mon maistre, vous aviés creu quelques choses au prejudice de l'entiere affection … 843/882 signs read as sense (95.6%), f. 243 98.3%, f. 241 92.5%. Open, all blocked from outside: two stretches of f. 241 under blots and a stain (no second copy, TNA has not digitised SP 106, DECODE has only 2300-px images), 14 name signs with no key, the sender's identity, recipient and year	targets/sp106r927/ · write-up
Lope de Soria → Charles V, Genoa, key-A letters (June and July 1523: R9488–R9490, R9492–R9493 body), RAH Salazar A-28 (9/28) (DECODE R9488–R9498; catalogue 149, class C)	Jun–Jul 1523	19 Sept 2026	partial	Split entry, part 1 of 2 (key B is under external plaintext). Key A had no key and no crib: about a thousand signs transcribed, then a simulated-annealing substitution solver on a 4-gram model of Castilian state papers of 1520–21 found the alphabet from random starts; the three copies of the June letter confirmed the homophones. Read nearly throughout: the Venetian league bought with Milanese money, Caracciolo to Venice, 18,000 ducats, the French king's expected descent on Italy, 5,000 infantry in Provence. Open: code words rip, pur, qed, mul	targets/soria1523/ · write-up
A Bavarian agent → the Dukes of Bavaria, report on the Pressburg talks, BayHStA Kurbayern Äußeres Archiv 4591 f. 263 (DECODE R9416 pp. 3–4; catalogue 161, B)	c. 1533–34	21 Sept 2026	partial	Split entry (the rest of the key volume is under adjacent plaintext). 3,150 signs, a homophonic substitution with ~23 frequent signs, broken by constrained annealing with the new de-1500s model on the first run; the faint gloss on the first lines agrees word for word. The Pressburg negotiations between Ferdinand and John Zápolya, a Turkish embassy, Imre Czibak offering the Sultan tribute, Alvise Gritti dispatched. Running German with local errors; person codes open	targets/kaa4591/ · write-up
A Vienna agent ("Sidon") → the Baron de Schenck, chamberlain of Augustus II, Vienna to Cracow/Lublin, fifteen letters, Archiwum Narodowe w Krakowie AKM 21 plik 30/1 (DECODE R7486–R7500, catalogued "NN z Drezna (jezuita?)"; catalogue 333, C)	7 Mar – 14 Jul 1706	9 Oct 2026	partial	Split entry, part 1 of 2 (the Sieniawski letters of the same DECODE group are under existing decipherment). German clear text with a 2-digit letter table and a 3-digit nomenclator. Free homophonic annealing failed; annealing only the block boundaries of an alphabetical table found it: a = 7 8 9, b = 10 11 12 … three to a letter, no j/v, 70 = w, 76–78 = y, 79 = z. 638 of 682 letter-cipher numbers read as sense (93.5%; "kein guttes omen", "der Baron Kleinburg alhier", "Reichsvicecantzler", "Malcontenten"); the recipient's own superscripts agree, and his gloss gives o. = der, 5 = von, 83 = der Kayser. Partial, 63.6% measured (865 of 1,361 cipher tokens, dotted word signs included; pass 2 of 9 Oct 2026): 65 of 77 three-digit codes open, nearly all persons with roles but no names (189 = Schenck, 150 = Augustus, 100 = imperial court, 204 = Sweden by context; the 500s are an alphabetical word list: 509 Friede, 513 Ordre, 515 Progressen, 519 Reise, 524 Urlaub); no key anywhere, the Wiener Diarium behind a bot check. Letters are dated Vienna, not Dresden.	targets/sieniawski1708/ · write-up
Emperor Ferdinand III (Vienna) → the Cardinal-Infante Ferdinand, Brussels, AGR Secrétairerie d'État allemande 540 (DECODE R1890, "Non-decrypted"; Schmeh Top 50 entry wrongly taken as Ernst's 2017 solve)	22 Feb 1640	9 Oct 2026	complete	Split entry, part 2 of 3; independent re-solution, not a first break (Andrew Aymeloglu, GitHub, 18 Sept 2026; his README, saying the 1640 additions are mostly syllables, was read before this work, his key and reading only after). The 1635 alphabet (part 1) covers 424 of 784 tokens; the rest is a syllabary in AEIOU series (40–44 = la–lu, 45–49 = na–nu, 50–54 = ra–ru, pla–plu = ta–tu, ap–up = ca–cu …), recovered by a family anneal on the la model, every top restart converging on the same table. All 782 tokens assigned (three copy slips, one obscured sign): Ferdinand asks the Cardinal-Infante for the 100,000 florins long requested, to be paid at Cologne, so that Hatzfeldt can raise troops against the Westphalian Circle's new armament	targets/ferdinand3/ · write-up

Key recovered based on plaintext from external sources

Key rebuilt by matching the ciphertext against a plaintext found elsewhere (an edition, a calendar, a copy in another archive).

Target	Date	Landed	Extent	Result	Where
Venetian ambassador at Rome → the Doge and Senate, ASVe IT ASVe 0045 010 Busta 27 f. 354 (DECODE R1862, "unknown sender (Roma) to unknown recipient", Non-decrypted, no transcription; catalogue 263, class C)	29 Apr 1628	22 Sept 2026	complete	Read: the cipher is R1874's, and the key was rebuilt from that record's contemporary decipherment. R1874's decipherment (~4,800 letters) aligned to its cipher by soft-EM gave a syllabary grid (prefix = vowel 6 a / 51 e / 62 i / 53 o / 64 u, last two digits = consonant), which reads this despatch unchanged. 1,280 groups transcribed here from the images (the record has none, and one manuscript line on page 3 had been missed entirely); 99.1% have a value and 0.95 of the text reads as sense (0.92 by a stricter window, 0.96 by hand audit). Cardinal Barberini asked twice whether Palma (Palmanova) was fortified and the Pontebba pass secured; the ambassador urged that the Pope go in person as legate to settle Mantua and Savoy. The word band proved alphabetical, which bracketed the open codes and corrected three values inherited from R1874 (628 sa, 6439 signor, 6481 i). Open: 12 word-code tokens (no other letter in this cipher survives to fix them) and one image-true line; the ambassador is not named.	targets/r1862/ · write-up
Speculum cipher flyleaf, anonymous alchemical note in Zentralbibliothek Zürich SCH R 201 (DECODE R4817; catalogue 271, class C)	Undated; volume 1618	23 Sept 2026	complete	Key recovered; one emendation. Rous’s partial reading (Korn 2023, p. 178 n. 41), found first and used as a crib, completed from the original image: PRIMA MATERIA, VOM WIDDER, OFFEN. All 79 signs assigned; 78/79 (98.73%) coherent without correction. Literal GEFAE retained; Gefäß conjectural. Ram gall, a hemispherical vessel and the furnace called the balance; annotation author/date unknown. Formerly non-decrypted, graphic signs, unknown language, one page.	targets/r4817/ · write-up
Armstrong → Madison, coded postscript ("THE = 972" code)	1808	14 Sept 2026	complete	48 of 49 groups determined, the 49th a probable slip for man; 580-group code table reconstructed from NARA pencil decodes. 2 Oct 2026: the table, extended by about 160 context values (additions972.tsv) and 21 frame-checked corrections to Founders, decoded the twelve other THE=972 despatches not printed decoded (Founders early-access texts, Dec 1807 – Oct 1808, and three of 1810 to Robert Smith on M34 roll 14): 15 Feb, 26 Mar, 31 May 1808 and the three of 1810 given a value throughout, 22 Feb 98.2%, 5/9 Mar 94.6%, 27/29 Dec 1807 88.9% (H/C share 39–79% per despatch); the code has punctuation groups (1105 comma, 962 full stop). 3 Oct 2026: the manuscripts found on M34 rolls 13–14 (with the 15 Nov 1807 triplicate, a 14th item) and the clerk's interlinear pencil decodes read at full resolution (~150 values, pencil.tsv); Founders' lists corrected from the frames; four of the five conflicts settled by third occurrences (946 on, 396 while, 1260 dou, 1415 ble/be). Measured as sense (value known or forced by its sentence, and the word fits): 13 of 14 items at 95 % or more, 1,553 of 1,587 groups (97.9 %): 15 Nov 1807 98.3 %, 27/29 Dec 99.6 %, 15 Feb 97.1 %, 22 Feb 96.1 %, 5/9 Mar 99.3 %, the postscript 97.9 %, 17 Feb 1810 95.2 %, the rest 96.6–100 %; 26 Mar 1808 (5 groups) 80 %, one digit doubtful. Open: ~25 single groups with no pencil over them, the 720/750 digit, the 1804–07 printed decodes not re-checked	targets/armstrong/ · write-up
Fra Giovanni di Lucca → Ferdinand III, 30 May 1644 (DECODE R2159, ÖStA HHStA)	1644	16 Sept 2026	complete	Read in full: a 24-figure Italian alphabet found by annealing from the eight consistent crib letters, all seeds agreeing and shuffled controls failing; Viterbi reads the 231 groups end to end. Checked against the DECODE images 22 Sept 2026: the apparent polyphony of 17 (i/n) was ten 14s misread as 17 in the published transcription, so only 19 = t/s is polyphonic; the clear text names the speaker, Grand Hetman Koniecpolski, at Brody. An offer to keep the Turk off Rákóczi, raise Moldavia against him and supply 2,000 Cossacks	targets/lucca/ · write-up
Feuquières → Catinat, Pignerol (418-group two-part "petit chiffre")	1691	16 Sept 2026	complete	(16 Sept 2026): the code is the petit chiffre of the Pignerol governors, companion of Bazeries' Grand Chiffre de 1691 (table transcribed from Gallica and verified on four 1690-91 despatches). A second letter in the same code, Louvois to d'Herleville of 6 Sept 1690, sits in the 1819 Mémoires t. I with its contemporary translation; the two letters share 72 groups, and hand alignment carried between them reads 586 of 601 tokens. Feuquières fixes the Veillane surprise: ready on receipt of Catinat's letter of the 24th, the enemy will not defend seriously, two roads into Avigliana (the faubourg and the houses along the pond), 80 horse to Saint-Ambroise, attack at two points, keep the dragoons from escaping. First reading since Bazeries' unpublished one of 1893	targets/feuquieres/ · write-up
Burghley (Westminster) → Walsingham (Paris), letter-book copy, BL Harley MS 260 f. 419r–v (DECODE R8364, catalogued as "Unknown sender to ...", Feb 1572; catalogue 122, class C)	29 Jan 1573 (Feb 1572 o.s.)	22 Sept 2026	partial	(6 of 26 cipher tokens, grade M). The six-sign place name "a Shipp sent to 2 0 ‡ H 6 1" is the word Digges 1655 prints in four letters of Nov 1572–Feb 1573 about the ship for the Duke of Alençon's escape (pp. 301, 303, 313, 327; also R8363): six signs every time, constant 1st and 3rd, reads VALERY, the St Valery of a French letter in Walsingham's file asking for a ship to be kept waiting (Cotton Vespasian F VI f. 197, 11 Dec 1572). About 45 cipher passages located in Digges. 1 Oct 2026: a Folger chronology (after Read 1960, pp. 94–95), pointed out by Adam Wright, summarises this letter under 29 Jan 1573: Davison = Maisonfleur, Hall = the Queen, Glasgow = Alençon (not the Archbishop; corrected here and on R8362), [9] and tt read as Alençon, the last gentleman = La Mole (grade I, by sense). Open: a nine-sign run and person codes 3, II, tt; the 1571 key (TNA SP 106/2 f. 141A, per Tu 2012) and the sent original (Cotton Vespasian F VI, not digitised) are not online.	targets/r8364/ · write-up
Giuliano Caprile (Eger, Buda) → Bonaventura Pistofilo / Alfonso I d'Este, and Alfonso Cistarelli (Eger) → Ippolito I d'Este, ASMo Ambasciatori Ungheria b. 4 (DECODE R1128–R1139; catalogue items 157 and 158, class B)	Mar 1519 – Apr 1521	21 Sept 2026	partial	R1129–R1135 read at the time or in 1882 (clear copies "b"/"c", interlinear glosses; Vestigia), R1138 deciphered by W. Somogyi (Scriptorium VI, 2025). Her two texts used as a crib to rebuild the 1520–21 homophonic sign cipher by EM alignment (714 letters, 90 % consistent); R1139 (6 Apr 1521) and R1136 (8 Mar 1520), with no decipherment anywhere, deciphered in part (custode and governor of Eger, the late bishop's money). (2 Oct 2026) No. 12 of 25 Feb 1521 (Vestigia 1921, not on DECODE) deciphered, 161 of 166 signs (97.0 %): the custode blamed for the claim that Eger was given away before the new bishop came; five key values added (lone o = h/u, oo = b, crossed z/x = q); second transcriptions at full resolution bring R1139 to 88.7 % and R1136 to 66.9 % of decoded letters; 3 Oct 2026 (passes 5–6, look-alike table and pen shapes): R1139 89.7 % ("mi per Milan"), R1136 unchanged; V1921 at 97.0 % counts as deciphered in full. R1128 (1519 two-tier cipher, Caprile's lawsuit) not read; Cistarelli's figure cipher R1137 gives fragments only to a ciphertext-only anneal.	targets/caprile1519/ · write-up
Sir Francis Walsingham → Edward Wotton, ambassador in Scotland, four letters, BL Add MS 32657 ff. 125–127, 140–141, 163–164, 194–195 (DECODE R4838–R4840, R4844, "non-decrypted; alphabet, graphic signs, numerical"; catalogue item 82, class C)	28 Jul – 10 Sep 1585	21 Sept 2026	partial	Signed letters in clear English with a two-digit name code. 28 Jul and 7 Aug are CSP Scotland viii nos. 40 and 65; 21 Aug and 10 Sep, uncalendared, transcribed here. The calendar's glosses fix 19 Arran, 39 Gray, 40 Justice Clerk, 24 Farnihurst, 20 Morton (Maxwell). Open: about fourteen sign-alphabet words on 10 Sep (the interlinear decipherment is blotted out; no key found) and numbers 13, 14, 23, 30, 37	targets/walsingham1585/ · write-up
Hugo de Moncada (Monaco) → Charles V, BNE MSS/20213/12 (BNE Digital "Carta 9"; DECODE R1191, filed without author as "Decrypted"; catalogue 273, class C)	6 Oct 1524	22 Sept 2026	complete	Solved earlier by Printed 1854 from the secretariat's CLARO: Colección de documentos inéditos XXIV pp. 417–419 (the pencil note "Descifrada en t.° 24 Col. Doc. inéd." on f. 1r; the volume's OCR reads "Moneada", which is why Tomokiyo and the catalogue missed it). The print has a gap ("que le llevase .... muchas razones"). 442 cipher signs transcribed here and aligned with the print by EM: homophonic letter substitution, a double cross after a sign makes a new sign, code groups lay = yo, ne = duque. The rebuilt key reads the gap: me rogaua que le lleuase a España, diziendome muchas razones para ello (Bourbon asked to be carried to Spain in the galleys). Open: three unprinted code groups (=D·, P‡D, =P‡D·f) and a line-end 99; 428/442 signs read (96.8%).	targets/moncada1524/ · write-up
Adrian of Utrecht, the Admiral and the Constable of Castile → Charles V, Vitoria, 30 Dec 1521, AGS Estado leg. 8 no. 150 (PARES: "cifrada prácticamente en su totalidad")	1521	18 Sept 2026	complete	Solved earlier by Deciphered by Claudio Pérez Gredilla, printed by Danvila, Memorial Histórico Español 38 (1899), pp. 703–705, with about 25 groups left unread; found by OCR search of the archive.org scan once PARES gave date and place. Here: the ciphertext transcribed from the PARES images and aligned with the print (19 xif = V. M. anchors), rebuilding a mixed nomenclator (trigraph words and syllables, homophonic signs, code initials in plaintext-alphabet blocks). The dead king is Manuel I of Portugal (d. 13 Dec 1521), not the print's "Rey de Inglaterra". About 17 unread groups read or corrected (rol agora, xep in cu-ya, dopca urcas, for como, fap Consejo; probable rip alianza, xim/xin voluntad, fir contra, pij manera), and the misprints "xiu", "itb", "xap", "40" shown to be xim, tib, xac and the group yo. About ten groups open (suf, xub, yo, ren nod, xac, sif, tib, xit, Lag, ne).	targets/adrian1521/ · write-up
Armstrong → Madison, 20 Feb 1808 (unique code, 369 groups to 1900, plus shorthand-like symbols)	1808	14 Sept 2026	complete	Adjudication of the AFIO contest solution (May 2025): the 56-entry key covers 36 % of the groups, leaves 93 of its own 133 mapped occurrences unread, rests on five numbers absent from the letter, and fits its sentence worse than a key built by the same procedure on a shuffled ciphertext. Madison himself wrote in May 1808 that no such cipher was in the office, and the roll-13 pencil decodes are all THE=972, so they cannot supply it. The claim does not hold; the letter stays unsolved	targets/armstrong/
Lope de Soria → Charles V (and Gattinara), Genoa, key-B letters (20 Jul – 13 Aug 1523: R9491, R9492 post data, R9494–R9498), RAH Salazar A-28 (9/28) (catalogue 149, class C)	1523	19 Sept 2026	partial	Split entry, part 2 of 2 (key A is under ciphertext-only). Key B (from late July) was rebuilt by aligning Soria's letter of 30 Dec 1523 (DECODE R9844) with the court's decipherment; the key-B letters read in substance: Siena to be governed at the Emperor's hand, Beaurain the bearer, 28,000 escudos for Swiss to descend on France, Andrea Doria discontented with the King of France. Open: about a dozen key-B groups not attested in the crib	targets/soria1523/ · write-up
Girolamo Ghinucci, Bishop of Worcester → Cardinal Wolsey, duplicate of 22 July 1527, BL Cotton MS Vespasian C IV f. 363 (old f. 332) (DECODE R8613, "Non-decrypted", dated 1529; L&P iv 3291 no. 2, "undeciphered"; tracker item 13)	22 Jul 1527	9 Oct 2026	complete	Syllabic cipher: consonant-row signs with superscript syllable digits, bare signs as letters, code words (L8 Regia Majestas, K5 Rex Gallorum, n2 Caesar, y6 Pontifex) and many nulls. Key rebuilt from three sibling letters of Ghinucci whose decipherments are in print or interlined: R8589 (L&P iv 5282 prints the Latin), R8590 (interlinear), R8476 (State Papers VI 552–556), and extended in R8613 by context. 797 of 814 cipher signs (97.9%) give Latin: the Imperialists hope to part Henry VIII from Francis I; three millions offered through Avemaria; the friar's report of Lautrec's order not to spare money for the French king's sons; the rumour that Wolsey goes to France to separate the English and French churches from Rome ad perpetuam divisionem. Agrees with L&P's English summary of the damaged original (Vit. B IX f. 127). Open: 17 single-occurrence signs. Gilino's letter in the same volume is a different cipher (own row, own page)	targets/ghinucci1527/ · write-up

Key recovered based on adjacent plaintext

Key rebuilt from plaintext beside the ciphertext: a partial interlinear or marginal decipherment, clear passages, a deciphered copy in the same file.

Target	Date	Landed	Extent	Result	Where
Ernestus dux and Joachimus; probably Ernst of Brunswick-Lüneburg and Joachim von Maltzan to Francis I, King of France, Eve of Saint Paul, at five in the morning; no year. Context suggests c.1518–1520;1519 remains (Dupuy 468)	Eve of Saint Paul, at five in the morning; no year. Context suggests c.1518–1520;1519 remains	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/3 pages transcribed; 1 more in part; 2/2 cipher pages deciphered. 684 elements, 660 identified (96.5%). Spot-checked here 13 Oct 2026: confirmed.	targets/dupuy468/ · write-up
Vicomte de Turenne to The Grand Master, 1 June 1528 (Français 2999 — Turenne, ff13r–13v)	1 June 1528	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/2 pages transcribed; 1/2 pages deciphered; 1 more in part. 785 elements, 775 identified (98.7%). Spot-checked here 13 Oct 2026: confirmed.	targets/turenne1528/ · write-up
Charles IX to Jean de Vulcob, 7 March 1573 (Français 3318)	7 March 1573	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: Transcribed; Deciphered. 754 elements, 746 identified; 4 groups with unresolved boundaries and 19 null-marked groups. Spot-checked here 13 Oct 2026: confirmed.	targets/vulcob1573/ · write-up
Scipion de Gonzague (Français 4696); not established for Français 4698 to Duke of Nevers (Français 4696); not established for Français 4698, 1585–1589 (Français 4696); 1590 (Français 4698) (Français 4696 / 4698 — graphical cipher, ff5r/62r–v, ff30r– 31r and f104r)	1585–1589 (Français 4696); 1590 (Français 4698)	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/7 pages transcribed; 5 more in part; 7/7 pages deciphered in part. 3,078 recorded groups, 1,147 identified; 8 further groups with unresolved boundaries. Spot-checked here 13 Oct 2026: partly confirmed.	targets/scipion1585/ · write-up
Dampierre to Duke of Nevers, 5 November 1589 (Français 3414 — Dampierre, ff125r–125v)	5 November 1589	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/2 pages transcribed; Deciphered. 67 elements, 67 identified (100.0%). Spot-checked here 13 Oct 2026: confirmed.	targets/dampierre1589/ · write-up
Not established to Not established, 1589? (Français 4696 — numerical cipher, ff54r–55r)	1589?	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 3/3 pages transcribed; 1/3 pages deciphered; 2 more in part. 1,744 elements, 1,692 identified; 1 obscured group of unknown length. Spot-checked here 13 Oct 2026: partly confirmed.	targets/fr4696num/ · write-up
Girolamo Gondi to Duke of Nevers and Rethel, 15 May 1594 (Français 4700 — Gondi symbol cipher, ff9r–10r)	15 May 1594	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 5/6 pages transcribed; 1 more in part; 3/6 pages deciphered. 1,469 elements, 1,449 identified (98.6%). Spot-checked here 13 Oct 2026: confirmed.	targets/gondi1594/ · write-up
Francesco Guicciardini? to Ferdinando I de Medici, Grand Duke of Tuscany, 10 May 1597 (Dupuy 155)	10 May 1597	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/9 pages transcribed; 1 more in part; 2/9 pages deciphered; 1 more in part. Transcribed passages: 1,952 elements, 1,946 identified. Spot-checked here 13 Oct 2026: confirmed.	targets/dupuy155/ · write-up
Brulart de Berny to Pierre Jeannin, 26 July 1607 (Français 3516 — Brulart de Berny, f132r)	26 July 1607	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: Transcribed; Deciphered. 266 elements, 255 identified (95.9%). Spot-checked here 13 Oct 2026: confirmed.	targets/berny1607/ · write-up
Villeroy; Henri IV to Pierre Jeannin and fellow envoys, 1607 (Français 3516 — Villeroy / royal cipher)	1607	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/15 pages transcribed; 2 more in part; 2/15 pages deciphered; 2 more in part. Transcribed passages: 1,145 elements, 1,099 identified. Spot-checked here 13 Oct 2026: confirmed.	targets/jeannin1607/ · write-up
La Boderie to Pierre Jeannin, 7 June1608 (Français 3516 — La Boderie, f59r–60v)	7 June1608	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 4/4 pages transcribed; 3/4 cipher pages deciphered; 1 more in part. 1023 elements,923 identified (90.2%). Spot-checked here 13 Oct 2026: confirmed.	targets/laboderie1608/ · write-up
Raimond Phélypeaux d’Herbault to Bullion, 29 July 1625 (Français 3944)	29 July 1625	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 1/2 pages transcribed; 1 more in part; 2/2 cipher pages deciphered. 375 elements, 366 identified (97.6%). Spot-checked here 13 Oct 2026: confirmed.	targets/herbault1625/ · write-up
Louis XIII to Philippe de Béthune, 30 June 1629 (Français 3675 — Louis XIII à Béthune, 30 juin 1629)	30 June 1629	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 1/2 pages transcribed; 1 more in part; Partly deciphered. 177 elements, 166 identified (93.8%). Spot-checked here 13 Oct 2026: confirmed.	targets/bethune1629/ · write-up
Châtillon to Servien, 3 août 1635 (Français 3758 — Châtillon, 3 August 1635)	3 août 1635	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 10/10 pages transcribed; 4/4 cipher pages deciphered. 666 written positions, 624 identified (93.7%); one group boundary uncertain. Spot-checked here 13 Oct 2026: partly confirmed.	targets/chatillon1635/ · write-up
Louis XIII to Châtillon and Brézé, 30 June 1635 (Français 3758 — Louis XIII, 30 juin 1635)	30 June 1635	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 10/10 pages transcribed; 6/6 cipher pages deciphered in part. 967 elements, 722 identified (74.7%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/louis1635/ · write-up
Brienne to Queen of Poland, Compiègne, 19 May 1646 (Clairambault 1067)	Compiègne, 19 May 1646	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 6/6 pages transcribed; 4/4 cipher pages deciphered. 350 elements, 349 identified (99.7%). Spot-checked here 13 Oct 2026: confirmed.	targets/brienne1067/ · write-up
Darnétal (catalogue attribution for manuscript p103; not every unsigned letter established) to Unestablished for the checked selections, Bordeaux, 8 September 1650 for p103; dates of the other selected letters unestablished (Clairambault 460 — reciprocal alphabet)	Bordeaux, 8 September 1650 for p103; dates of the other selected letters unestablished	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 5/5 pages transcribed; 5/5 pages deciphered. 888 visible elements, 887 identified; additional ciphertext hidden in the binding. Spot-checked here 13 Oct 2026: partly confirmed.	targets/clair460alpha/ · write-up
Not established to Duke of Bouillon, 6–8 August 1650, Madrid (Clairambault 460 — graphic cipher, pp53–55)	6–8 August 1650, Madrid	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 3/3 pages transcribed; 3/3 pages deciphered. 949 elements, 941 identified (99.2%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/bouillon1650/ · write-up
de Loménie Brienne to Servien, 10 January and 1 August 1653; letter B undated in selected cipher pages (Français 5160)	10 January and 1 August 1653; letter B undated in selected cipher pages	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 14/14 pages transcribed; 13/13 cipher pages deciphered. 1,356 elements, 1,308 identified (96.5%); 40 nulls ?, 3 unidentified, 5 cancelled. Spot-checked here 13 Oct 2026: confirmed.	targets/brienne5160/ · write-up
Louis XIV to Marquis de Feuquières, 9 December 1678 (NAF 22738)	9 December 1678	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 4/4 pages transcribed; 4/4 pages deciphered in part. 1,920 elements, 1,169 identified (60.9%). Spot-checked here 13 Oct 2026: confirmed.	targets/feuquieres1678/ · write-up
Abraham Duquesne to Seignelay, 25 September 1681; later letter in the same September correspondence (NAF 22736 — Duquesne, ff137r–138r and144v)	25 September 1681; later letter in the same September correspondence	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 4/4 pages transcribed; 4/4 pages deciphered. 751 elements, 743 identified (98.9%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/duquesne1681/ · write-up
Jacques Vergier to Monseigneur, probably Jérôme de Pontchartrain, 10, 20, 26, 29 and 30 March 1696; 6 April 1696? (Clairambault 1108)	10, 20, 26, 29 and 30 March 1696; 6 April 1696?	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 17/17 pages transcribed; 12/12 cipher pages deciphered. 1,212 elements, 1,209 identified (99.8%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/vergier1696/ · write-up
Paget? to Votre Excellence; identity unestablished, Genoa, 3 April and 28 August 1714 (Clairambault 1225)	Genoa, 3 April and 28 August 1714	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 13/13 pages transcribed; 7/7 cipher pages deciphered. 495 elements, 487 identified (98.4%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/clair1225/ · write-up
De Villardeau (Holstein letter, signed on p125); testament memorandum anonymous to Unestablished, Undated selected sheets; Russian-correspondence dossier, 1725–1730 (Clairambault 531 — testament and Holstein marriage)	Undated selected sheets; Russian-correspondence dossier, 1725–1730	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 6/6 pages transcribed; 5/5 cipher pages deciphered. 1,248 elements, 1,237 identified (99.1%). Spot-checked here 13 Oct 2026: confirmed.	targets/villardeau/ · write-up
F. Le Rouge to the Chancellor, 11 July, year unestablished (Français 2971)	11 July, year unestablished	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: Transcribed; Deciphered. 623 elements, 610 identified (97.9%). Spot-checked here 13 Oct 2026: confirmed.	targets/lerouge/ · write-up
M. de Villeroy to M. de La Verrière, 19 July; year not written (Dupuy 61 — Villeroy, f258r)	19 July; year not written	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/2 pages transcribed; Deciphered. 1164 elements, 1147 identified (98.5%). Spot-checked here 13 Oct 2026: confirmed.	targets/villeroy258/ · write-up
Not established to Not established, 1650–1653 volume; individual dates unchecked (Clairambault 460 — numerical cipher, pp289–303)	1650–1653 volume; individual dates unchecked	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 9/9 pages transcribed; 9/9 pages deciphered in part. 2,080 visible elements, 1,558 identified; additional ciphertext hidden or damaged. Spot-checked here 13 Oct 2026: partly confirmed.	targets/clair460num/ · write-up
Several correspondents; Chamilly and Barrière named in headings to The Prince in several headings; another damaged heading names a duke, 18, 19 and 26 August; year not visible in this source extract (Clairambault 460 — barred letters and numerals, pp285– 287)	18, 19 and 26 August; year not visible in this source extract	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 3/3 pages transcribed; 3/3 pages deciphered. 1,218 visible elements, 1,139 identified; additional ciphertext hidden or damaged. Spot-checked here 13 Oct 2026: partly confirmed.	targets/clair460bar/ · write-up
Signed DC to Not established, 20 May; year not established (Français 3321)	20 May; year not established	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: Transcribed; Deciphered. 49 elements, 46 identified (93.9%); 2 ?, 1 unidentified. Spot-checked here 13 Oct 2026: confirmed.	targets/fr3321dc/ · write-up
Unidentified to Unidentified, Unestablished;1554/1555 remains a hypothesis (Clairambault 1161)	Unestablished;1554/1555 remains a hypothesis	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 6/6 pages transcribed; 2/6 pages deciphered; 4 more in part. 3,149 elements, 3,046 identified; 5 further wordlike groups not classified. Spot-checked here 13 Oct 2026: partly confirmed.	targets/clair1161/ · write-up
unknown to Bertelot (catalogue item73), undated (Français 3822)	unknown	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/5 pages transcribed; 1/5 pages deciphered. 1,646 elements, 686 identified; 10 further groups with unresolved boundaries. Spot-checked here 13 Oct 2026: partly confirmed.	targets/bertelot/ · write-up
Joachim de Dinteville (Langres) → the Duke of Nevers, BnF Français 3621 no. 116 f. 130 (DECODE R9451, dated there 3 July), with its sibling no. 114 f. 128; catalogue 183, A	4 July 1592	7 Oct 2026	complete	Key rebuilt from the interlinear decipherment of f. 128 by an outside contributor, Alex Zarco (PR #24), completing NoAutopilot's published partial key; verified here. Homophonic substitution of letters, digits and signs whose crosses and strokes make separate signs (∇ a / ∇ with stroke t), four nulls; the 21 Sept attempt failed because the marks were merged and the gloss misread (Gennes, not Geneve). f. 130: 514 of 522 cipher tokens read (98.5%), shuffled keys best −4.29 against −2.00; f. 128 agrees with its 1882 print (Revue de Champagne XII p. 340). Dinteville alone with the townsmen at Langres, no taille without troops, asks leave to withdraw. Eight unclear signs open	targets/dinteville1592/ · write-up
Coenraad van Beuningen → Gaspar Fagel, London, Nationaal Archief 3.01.18 inv. 249 scans 131–132	23 May/2 Jun 1676	5 Oct 2026	partial	Key recovered based on adjacent plaintext. No earlier reading found in the sources checked; no contemporary decipherment of this letter; key rebuilt from the clerks’ sibling glosses in inv. 248 (1675), 249 (1676), 250 (1677). Strict H+C+M 264/278 (94.96%) with all glosses and gemelte; 263/278 (94.60%) excluding later hands; 265/278 (95.32%) only with glossed selve and later-hand genomen. Same-code direct gloss support 241/278 (86.7%). Du Plat, bylanders at Bruges, the French dessein on Den Briel and Hellevoetsluis, and a doubtful informer. Twelve residual I tokens, g121 if gemelte is retained, and unread s439; unglossed siblings remain workable. Blind test M 10/10, I 3/6, not cleanly blind. Shuffle: 93% versus 1.2% mean, 11.6% maximum over 1,000 keys. Feyseel Nur (with Claude and Codex); both reviews retained.	targets/beuningen1676/ · write-up
François van Aerssen → Johan van Oldenbarnevelt, Paris, Nationaal Archief 3.01.14 inv. 2019 scans 57 and 66	25 May; 2 Jun 1601	4 Oct 2026	partial	Key rebuilt from the 11 and 28 May interlinear glosses and 30 April list A (scan 46); decipherment by Feyseel Nur (with Claude and Codex). No contemporary decipherment of the target passages; unprinted in the editions searched. 25 May: 59/61 tokens (96.7%) attested at least once, 58/61 (95.1%) at least twice after qF/qT and plain-42 corrections. Open: 47' unidentified, 45' conjecturally Essex (I). Peace feelers and Neville in the Tower; Winwood is contextual identification only. 2 June sibling: 9/11 (81.8%), below bar; combined 68/72 (94.4%). List A comparison 281/330 (85.2%; old frequency-H band 92.0%); now reused as attestation, not a clean holdout. Both Codex reviews retained; coverage is not accuracy. LLM transcriptions from Nationaal Archief scans checked against the glosses and list A used to reconstruct the key; qF (foot bar) versus qT (stroke through descender) checked visually on scans 43, 57 and the 11 May scan.	targets/aerssen1601/ · write-up
Snell → an unidentified correspondent ("A?ud"), invisible-ink message copied for the Hessian authorities, HStAM 9 a Nr. 259 f. 249 (HCPortal record 513, "polyalphabetic", not solved; catalogue 340, C)	20 Feb 1824	28 Sept 2026	complete	Key rebuilt from the decipherer's note on the leaf; whole text read. The note gives the key letters bcdefg and the first word; the table is a Vigenère without j (b..g = +2..+7) whose columns run on past z into the signs 1–4 ("de 4ette" = de Wette). 169 signs, one omitted sign and 11 copy slips corrected. Questions about Paul Follen, America, Fries, Oken, de Wette and "the discovery of this plan in Wezlar". Addressee and Lisching unidentified	targets/hesse1824/ · write-up
John Quincy Adams → Secretary of State, St Petersburg, No. 88, 25 June 1812 (Ford, Writings 4:357, "[Nine lines of this paragraph not decyphered.]"), NARA M35 reel 3 frames 0206–0207	1812	22 Sept 2026	complete	Solved: premise refuted, code rebuilt, read (22 Sept 2026, third pass: 380 of 390 groups read as sense, 0.974, measured; Ford's nine lines 133/133, the remainder 247/257 = 0.961). Adams's despatches are not in Armstrong's THE = 972 code, as the Madison editors state: every coded despatch on M35 reel 3 (and reel 2 frame 0155) has the = 1385, of = 1576, that = 1384, and = 668; he held Armstrong's cypher only to write to Paris (Ford 3:328, 3:370). His own code rebuilt to 1,166 numbers from the clerk's interlinear decodes (frames 0047, 0048, 0182–0185, 0189, 0190, 0194, 0198, 0206, 0210–0211, 0247) and Ford's printed decipherments of 1811–1812 blocks (No. 95 aligned against the Department's clear copy). The nine lines: the journey to Wilna went against the Chancellor's advice, Rumyantsev's apoplectic stroke soon after reaching Wilna, a second and worse one before 25 June, a false report of his death in the German gazettes. The remainder (frame 0207): Kurakin's note on the evacuation of Prussia, which Lauriston called a proposal "France had never thought of making to Russia even after the battle of Friedland"; Napoleon expected to strike at the centre of the empire rather than along the shores of the Baltic. Open: 7 groups cut by the binding gutter on the film (needs the original), 3 single code groups (Adams's copying slips 1112 1433 read as intended, 'impaired')	targets/jqa/ · write-up
Court informant → Louis de Gonzague, duc de Nevers, Paris, BnF fr. 3976 f. 62 (DECODE R3705; sibling ff. 131, 133 same informant; catalogue item 187, class C)	29 Apr 1588	22 Sept 2026	complete	(350 of 353 plaintext letters of the 19 cipher runs, measured) with a key rebuilt from the interlinear glosses of the same informant's letter of 4 June 1588 (f. 131): Henri III 'cuyda estre resolu au conseil de prendre et jetter en l'eau cinq ou six des plus grans ligueurs de' Paris, urged by the Queen, the chancellor and d'O, held back by Villeroy's remonstrances; Guise's answer that he had '40 mil hommes'. The name in run 1 (Fougères?) and code numbers 28, 49, 70 open (10 = Épernon, glossed on f. 139). The key's source letter, f. 131 (4 June 1588), was itself deciphered in full on 2 Oct 2026 with the key: 130 of 130 cipher signs and 8 of 8 code numbers (9 Guise, 10 Épernon, 11 le Roy, 24 Nevers, 72 Savoie, 73 ambassadeur), every run agreeing with its gloss of the time (G04 is "que s'il perissoit", not "pouvoit"): the Hôtel de Ville assembly moving to an apology to Henri III, the Brissac/Elbeuf rumours at Rouen, Épernon at Loches, and Savoy arming to take the marquisate of Saluzzo.	targets/nevers1588/ · write-up
Thomas Spinelly (Brussels) → Henry VIII, BL Cotton MS Galba B V ff. 40–41 (DECODE R8416, "Spynelly to unknown recipient, London 1516, non-decrypted"; catalogue item 100, class C)	1 Feb 1517	21 Sept 2026	complete	(21 Sept 2026). A clear despatch with five lines in a substitution of invented signs, no decipherment; broken from the crib "audensier" in the clear text. The Audiencier and Luigi Marliano expect the King's entry into Brussels at the end of the week, the Emperor to go to Cambrai about the 15th (L&P II 2862 summarised only the Cambrai clause). Word signs for the Emperor and (probably) the King read from context.	targets/spinelly1516/ · write-up
Francis Needham (before Sluys) → [Walsingham], BL Harley MS 287 ff. 39–40 (DECODE R8479, "Needham to unknown recipient"; catalogue item 133, class A)	28 Jul 1587	21 Sept 2026	complete	(21 Sept 2026). Three-grid pigpen (nine box shapes plain / dot below / dot inside, alphabet in order), the Wilkes–Walsingham cipher; key rebuilt from a few contemporary glosses; every bare run of f. 39v read: no intelligence of the enemy, whether there was a bridge, the channel closed with hoys and flyboats chained and backed with flat boats. One word open.	targets/needham1587/ · write-up
Richelieu → M. de Rancé, BnF Français 3829 ff. 87 & 89	1629	14 Sept 2026	complete	Homophonic alphabet and nomenclature recovered ciphertext-only; later matched word for word to Avenel (1858), which the catalogues missed	targets/richelieu/ · write-up
Huang Xing → Lin Hu and Li Genyuan (JACAR B03050731500)	1916	15 Sept 2026	partial	Scheme identified: three kana per character, consonant row carries the digit, vowel free; plaintext read from the JACAR frames	targets/sunyatsen/HUANG_NOTES.md · write-up
Maltravers → Ormonde	1634–35	15 Sept 2026	complete	Regular block alphabet and nulls recovered from 59 figures; every spelled word reads; the nomenclator then confirmed clause for clause against Wentworth's dispatches in Knowler's Strafforde's Letters (1739). Two person-codes in one clause remain unidentified	targets/ormonde/ · write-up
Lanssac → Charles IX, Warsaw, 26 Apr 1573, BnF fr. 4735 no. 51 f. 124, and the siblings ff. 154, 156, 160, 164, 174 (catalogue item 3, class A)	1573	17 Sept 2026	complete	(17 Sept 2026): both cipher passages of f. 124 (96/100 signs) from the Gallica images; homophonic letter cipher with word signs, key fixed from the glossed sibling f. 154v ("car je n'ay pas cinquante escuz"), the gutter-cut marginal fragments, and a pinned annealer with shuffled and blind controls. The key then reads the three election letters whose "déchiffrement" is only that cut gloss: 1 May to the King, ceste nation est autant vénale et sujette à se laisser gaigner par argent comme sont les Allemans; 1 May to Anjou, l'Empereur en y a despendu plus de trois cens mil … faisant le pis qu'il peult contre vous; 9 May from Płock, the election carried contre la volunté et menées du Grand Seigneur, de l'Empereur, des princes de l'Empire, du Roy d'Espaigne, du Moscovite et du Roy de Suède, qui tous estoient bandez contre vostre Majesté. Tomokiyo's table for the cipher corrected. (2 Oct 2026) The two long letters of 24 April re-transcribed at full scale and deciphered with the key, extended by about a dozen signs: to Anjou (f. 154r 96.7 %, 154v 92.0 %), the Lithuanian vote and the Sultan's recommendation made avec ruse; to Catherine (f. 156r 97.2 %, 156v 88.2 %), vous serez mere des deux plus grans roys du monde, Denmark and the Sound, money and galleys; 1,521 of 1,639 signs (92.8 %) at that pass. (3 Oct 2026) Every open group re-read at 2x against Tomokiyo's current table (st, il, nous, par, t adopted) and ff. 160, 164, 174 re-transcribed word-aligned: every cipher letter is over the read bar, measured per letter: f. 124 96 %, to Anjou 97.5 %, to Catherine 95.8 % (the cipher shows the German princes se monstrent estre ennemis jurez de … France, where the decipherer wrote "si monsieur estoit ennemy"), f. 160 100 %, f. 164 98.1 % (car si jusques cy … a despendu trente mil escus), f. 174 100 %; 2,094 of 2,159 signs (97.0 %). Open: 65 signs at the microfilm gutter or under overwriting (Gallica has one microfilm only), single signs, and one word sign on f. 164 that is in no key. The four 9 May letters ff. 182–189 are in clear.	targets/lanssac/ · write-up
García de Toledo → Philip II, Messina, 16 July 1565, AGS Estado leg. 1394 no. 247 (duplicado de carta cifrada; catalogue item 33)	1565	18 Sept 2026	complete	(18 Sept 2026): a homophonic substitution on two-digit figures 12–43, in plain alphabetical order (a 12–14, b 15, c 16 … u/v 38–40, x 41, y 42, z 43; vowels three figures each, consonants one, no nulls and no nomenclator), used inline in otherwise-clear Spanish; 1,930 figure groups read end to end from the PARES images. Two runs beside clear words broke it — seiscientos after "los", en tiera after "puso" — and the alphabetical order then predicted the unseen values (f 21, x 41, y 42, z 43), each confirmed on effetos, baxeles, yban, plazer. The viceroy of Sicily, a week after the Piccolo Soccorso got into Birgu, reports a second relief run waved off: the galleys came within four miles of the harbour mouth, the barqueta sent ahead arranged the fire-signals, and La Valette signalled clarissimas that they must not enter, twice within the appointed hour, so they slipped back out to sea undetected — as Mosiur de Lenni (Andrea Provana di Leinì) and the rest report, with Çanoguera piloting. Also: La Valette embracing Cardona's 600 soldiers and weeping, unable to speak; the forbidden sortie that killed 200 Turks; Turkish lines "seven, eight and ten palms high"; 1,500 Turks at the water and sickness killing men in the Ottoman camp; the Algiers squadron off Cabo Bono; 40,000 escudos a month for the ships, half the 48,000 ducados already eaten, and El estar sin galeras me destruye. The companion covering letter of the same day, AGS Estado leg. 1394 no. 249 (PARES 3576586, duplicado), was deciphered in full on 2 Oct 2026 with the key unchanged, 150 of 150 cipher tokens (two new signs: barred θ = ll, an overlined M-like sign = r): La Valette's latest letters are harto mas apretado than expected, the Algiers ships off Cabo Bono, the galleys for la Goleta held for the respuesta del Maestre and leaving mañana. Tomokiyo does not list it and PARES records no decipherment; the non-duplicate originals of 247 and 249 and the Mdina governor's enclosure are still to be found	targets/toledo1565/ · write-up
Marqués de los Balbases (Nijmegen) → Baltasar de Fuenmayor (Copenhagen), fourteen letters, 29 June 1677 – 22 Mar 1678, AGR Brussels SEG inv. 2559 (DECODE R985–R998, "partially decrypted"; catalogue item 40, class A)	1677–78	19 Sept 2026	complete	(19 Sept 2026). No key survives and none of the series keys R958–R965 fits. Ten letters carry a contemporary margin decipherment; four (19 Sept 1677, 1, 8 and 22 Mar 1678) carry none. The key was rebuilt from the ten margins (repeated-word anchors, then a hard-EM aligner): letters and CV syllables in numbers, letter pairs with disguised consonant (xo = co, no = fo), reversed vowel-first pairs (ef = se, ad = pa, ac = ma), ~20 word codes (d+ que, H V.S., des ministro, ban Flandes), nulls. It covers 97% of 8,150 groups and reads the four undeciphered letters: the Imperial protest over the Bremen assembly, the Brunswick claim to ambassador rank, the fall of Ghent. Open: rare name codes, a few blurred homophones	targets/balbases1677/ · write-up
The Hereditary Prince (Berlin) → Prince Frederik (filed with William V), KHA The Hague, Prins Willem V inv. 337 (DECODE R2239, "non-decrypted", "book cipher"; catalogue 218, class C)	12 Mar 1796	21 Sept 2026	complete	Read: a numbered word-list code rebuilt from the sibling R2237. Groups take a whole entry, chosen letters of it, or the entry cut by k letters at the front or back; R2237's bound contemporary decipherment gave about 110 entries, which read 70% of R2239. The body then proved to be printed in Colenbrander, Gedenkstukken II no. 752 (1906); used as a crib it completed the list, and the unprinted opening was read (the Haugwitz–Elgin conference of the 8th, Haugwitz at the King's ball). One word (entry 244) open. R2236 (Dutch, another list) not read.	targets/hereditary1796/ · write-up
Marqués del Carpio (Rome) → Baltasar de Fuenmayor (Copenhagen), ten letters, 4 Sept – 11 Dec 1677, AGR Brussels SEG inv. 2559 (DECODE R1002–R1011; R1004 and R1005 "non-decrypted"; catalogue item 38, class A)	1677	19 Sept 2026	complete	(19 Sept 2026). The syllabic nomenclator was rebuilt from the letters themselves. It has letters 9–30, consonant-vowel syllables 31–60 in rows of five, struck numbers for a second syllable table, letter pairs with disguised consonants (ge = de, bo = co, ma = la) and ~20 code words. It reads all ten letters, including the two DECODE lists as unread, whose faint marginal decipherments agree with the key reading. Content: the French fleet off Messina and Catania, Orange raising the siege of Charleroi, Innocent XI refusing Carpio audience, the fall of Freiburg, and d'Estrées at Turin threatening war in the Milanese. The eight margin letters are summarised, not edited; three code words are conjectural	targets/carpio1677/ · write-up
Jörg Weinmeister (Hungary) → the Dukes of Bavaria, King John's camp before Palota, spring 1533, BayHStA Kurbayern Äußeres Archiv 4591 ff. 232–235 (DECODE R9407, catalogued "Augurellio to …"; catalogue 164, class B)	late Apr–May 1533	22 Sept 2026	complete	(96.7% of 1,003 plaintext words, measured). "Augurellio" is only the cover; the writer signs in cipher twice, Iorg Weinmaister. System A′ of the volume, key carried over from R9410's gloss and extended; a second pass fetched the record's overlooked image P6 (f. 235r, 36 lines; the first pass's angled "f. 234 column") and read all four cipher pages line by line against the scans. Dated by its contents, not the docket ("10 Decembris 33", first read 35): King John's siege of László Móré in Palota (Łaski, Turks, Bohemian miners, from 14 Apr 1533), the diet of Pressburg, Török Bálint sent away from Ferdinand's table at Vienna to come back sober, Katzianer's threat, Simontornya taken after Easter, copies of the dukes' letters in Minckwitz's hands at Kraków, the Munich merchant Hans Schetinger's Vienna factor as the post, fifty thousand Turks marching from Belgrade on Buda. Hourglass sign = F.G., dotted ʃ = Ferdinand. Open: ~31 scattered words, the ∇ name sign (probably Łaski), one name	targets/augurelio1535/ · write-up
Sir Edward Stafford (Paris) → [Sir Francis Walsingham], BL Harley MS 1582 ff. 76–77 (DECODE R8504, catalogued as "Sr. Ed. Stafford to unknown recipient", London; catalogue 117, class C)	9 Nov 1586	22 Sept 2026	complete	Clear holograph letter; its one cipher word (7 signs) reads princes with the key rebuilt in harley1582r8500. Not in CSP Foreign.	targets/harley1582r8504/ · write-up
Sir Edward Stafford (Paris) → [Sir Francis Walsingham], BL Harley MS 1582 ff. 65–66 (DECODE R8500, catalogued as "... to unknown recipient", London; catalogue 119, class C)	20 Aug 1586	22 Sept 2026	complete	Five postscript cipher runs, never deciphered on the page, read with Stafford's letter key rebuilt from the f. 72r decipherment (R8501) and f. 73r glosses (R8502): Junius' son from Cambrai, his credit with Balagny; 6 = e, not a null. Four sign values by context.	targets/harley1582r8500/ · write-up
Paris informant (same as ff. 62, 131) → a member of the duc de Nevers's household, BnF fr. 3976 ff. 133–134 (DECODE R3708; catalogue item 186, class C)	6 Jun 1588	22 Sept 2026	complete	(226 of 230 cipher signs of the 18 runs, measured; 4 single-occurrence signs from context) with the key rebuilt for f. 62 (nevers1588), extended by three signs (ϙ = y, crossed 4 = z, bracket = n): Paris three weeks after the Barricades, the 'curez St André et St Benoist' who 'font rage de mutiner le peuple', 'le prevost Chapelle' carrying the Hôtel de Ville's decisions to 'le conseil de 9 [Guise] et 106' with 'l'evesque de Lion' and 'Nuilly', and the échevin 'nommé Cotteblanche' refusing to let 'l'artillerye de l'Arsenac' go to Corbeil without the King's leave. Also re-reads the f. 131 gloss over code 24 as Nevers (not Navarre). The informant's letter of 9 June (f. 139, not on DECODE) read with the same key, 106/106 signs, every run matching its contemporary gloss; its glosses give 10 = Épernon, 38 = Garrault, 98 = la paix, 107 = Grand Écuyer. Codes 4, 23, 25, 33, 79, 94, 105, 106 open; recipient unidentified.	targets/r3708/ · write-up
Edward Wotton, ambassador in Scotland → Sir Francis Walsingham, three despatches, BL Add MS 32657 ff. 167–168, 173–174, 175–176 (DECODE R4841–R4843, "non-decrypted, numerical"; catalogue item 81, class C)	25 Aug – 1 Sep 1585	21 Sept 2026	complete	Read — clear drafts, code names identified (21 Sept 2026). The records are Wotton's own drafts in clear English; persons and places are one- and two-digit code numbers. Read in full; 19 = Arran, 39 = Master of Gray, 10 = King James (certain); 24 = Kerr of Ferniehirst and 40 = Justice Clerk Bellenden (CSP Scotland viii, checked later the same day); 9 England/the Queen, 20 Maxwell, 3 France (probable); 27, 32, 36 unidentified. Gray's plan to seize the King and Arran, the prelude to Stirling (Nov 1585).	targets/wotton1585/ · write-up
Johan Banér → Torsten Stålhandske, Hof, intercepted copy in Kircher's correspondence, APUG 568 f. 239 (posted unsolved on Cipherbrain, 2 June 2019; new-solves catalogue entry "Intercepted letter, Banér to Stålhandske")	29 Dec 1640	22 Sept 2026	partial	Read in part: key rebuilt here from context cribs, 246 of 275 cipher tokens (89.5%) inside words read as sense, measured 5 Oct 2026. Homophonic two-digit alphabet (70 values) with 14 three-digit code words; none of the Swedish keys on DECODE fits (Banér's R4122, R4123/R4325, R4266 tried). Cribs durch and und let the annealer write nach der Oberpfaltz und gegen die [442]; the key was extended word by word from there: etwa auch die in Schlesien agirende [361] dazu ziehen, gute acht geben, Fürsichtigkeit nachvolgen, durch Espionen, wohin er seine Marche zu richten. Transcription checked against the Museo Galileo scans of f. 239 (four figures corrected). Open: the second courier's name and a stretch in S5, both confirmed on the manuscript as corrupt in the only copy; all 14 code words (no key). New shared model de-1640s.	targets/baner1640/ · write-up
Lope Hurtado de Mendoza (Rome) → Charles V, nine letters, Sept–Nov 1522, RAH Salazar 9/26 (DECODE R9634, R9644–R9646, R9648–R9650, R9652, R9656; catalogue 144, class B)	Sept–Nov 1522	20 Sept 2026	partial	Cipher broken and the key rebuilt; four letters read complete as to content. Bound among Alonso Sánchez's Venice letters and catalogued with them, but a different, unpublished cipher — Sánchez's groups end strictly in b c d f g h l m n, Hurtado's take finals p/z, and xul is despues in Sánchez against lo here; Tomokiyo's 2025 article covers Sánchez and Juan Manuel, not this. The crib was bound with it: R9644 and R9650 each carry the clerk's own full contemporary clear version beside the cipher, keyed by marginal letters A/B/C. 47 code values fixed against that plaintext, 29 probable (kept apart; an audit produced 4 retractions, and the 10 substitution-alphabet signs were later downgraded to probable when an 8× zoom showed signs that must be different letters looking near-identical at the magnification they were read at — the code groups, plain latin trigrams that cross-check, are unaffected). The key reads where it was not built: 70%/62%/57% of tokens in R9648/R9650/R9645, and el papa unaided in R9656, which has no crib. Structure not in the published account of the siblings: homophones at code level (ta/tu both no), codes carrying syllables (na·da, que·da·rian, con·ocia) and codes running together (porque, quales). Read complete: R9644, R9648 (its duplicate), R9650, R9652 — Adrian VI temporising between Charles V and Francis I, the French into Italy with a great sum to Lyon, Adorno's intercepted letters, the Pope tan escasso que nunca hombre lo fue mas. Correction to DECODE: R9652 carries no cipher at all. 2 Oct 2026: R9649 deciphered in full from its own clear (f. 268; the Ostia fortress, the licenciado Bernardino); the key extended on R9634, which is from Genoa, 13 Sept 1522 (DECODE has the month only): Geronimo Adorno's plan for Charles V and Henry VIII to invade France through Provence, la tierra mas flaca que el tiene, 70,000 ducats a month, set for March — 846/1191 tokens read as sense (71%); then R9645 73% (f. 243r = the Ferrara capitulation terms), R9646 80%, R9656 61%, all measured. 3 Oct 2026: codes pooled over all 21 Hurtado letters, spelled words matched against a Spanish lexicon, R9656 re-transcribed from the images: R9634 872/1191 (73.2%), R9645 75.3%, R9656 177/225 (78.7%), 76% over the five; still partial (DECODE has no larger image; the RAH Biblioteca Digital is behind a bot check). New values: ɋ = r, plural z after codes, yob ducado, xu gente, nulls after the opener Ꮒ. Open: about 40 single-occurrence code groups in R9634 (no crib, not in CSP Spain ii) and spelled words with one sign at the limit of DECODE's resolution in all four partial letters	targets/lopehurtado/ · write-up
Lope Hurtado de Mendoza (Rome) → Charles V and the Grand Chancellor Gattinara, eight letters, May 1523 – Mar 1524, RAH Salazar 9/28 = A-28 and 9/30 = A-30 (DECODE R9667, R9683, R9695, R9846, R9866–R9869; catalogue 148, class C)	May 1523 – Mar 1524	21 Sept 2026	partial	Catalogued as June 1523, unknown recipient: five letters to the Emperor and two to Gattinara, four of them in 9/30 and dated 1524. R9667 (5 May 1523) and R9846 (5 Feb 1524) were calendared by Bergenroth (CSP Spain ii 548, 617) from contemporary decipherings; R9667's bound "Claro" leaf is the Abbot of Nájera's, not Hurtado's. The 1522 cipher with a changed key (tef paz → nec-): 105 values (65 confirmed) rebuilt from the clerk's clear of R9846 on f. 133r. The six uncalendared letters read in part (R9867 ~60%, R9869 ~two thirds, R9866 a quarter to a third, R9868 f. 430r, R9683/R9695 fragments): Beaurain's mission, the Pope wanting France ruined or a truce, the Datary Giberti, Hurtado's unpaid salary. About 30 code groups and look-alike signs open.	targets/lopehurtado1523/ · write-up
Ferdinand Sigmund Kurz von Senftenau (Hamburg, Glückstadt) → Trauttmansdorff / Ferdinand III / Schwarzenberg, nine letters, SOA Plzeň, RA Trauttmansdorff inv. 200 kart. 9 (DECODE R3811–R3815, R4625, R4645, R4734, R4736; catalogue 67, class A)	Jan–Apr 1639	19 Sept 2026	partial	Key rebuilt; all nine partly deciphered. The interlinear glosses on R3811 p.1 give a regular table: 41–100 are consonant+vowel signs, two consonants per ten in reverse alphabetical order (41 r, 46 p, 51 n … 91 b), units o i e a, plus a separate -u series in the fifth places (45 pu … 100 ru); 1–40 and ~20 letter/graphic signs are letter homophones (a few for pairs: 9 te, 6 se). 73 images transcribed and checked against the images in a second pass (21 Sept 2026: ~360 signs corrected, 56 skipped lines restored; 23,699 signs, 99.9% keyed); structured annealer + German LM. Content: Banér in Lower Saxony, Arnim's pardon, the Danish mediation and its expiry, "ein teufflisch inventum" of universal treaties. Mírka (2012) had rebuilt the key and read R3811 only, printing neither. Open: residual garble where the image does not decide the digits, and the Latin passages; R3812 is 1639, not 1638	targets/kurtz1639/ · write-up
Nicolò Sadoleto → Ercole I d'Este (and a copy of his letter to King Ferrante), Pozsony, 15 Jul – 27 Nov 1482, ASMo Ambasciatori Ungheria b. 1/9 nos. 7, 8, 11, 26 (DECODE R1101, R1102, R1103, R1106; catalogue 154, class A, pre-1497)	1482	19 Sept 2026	partial	Read in part; two of four found read at the time. Never printed (MDE skipped Sadoleto). The sign cipher (monoalphabetic, homophones for vowels, n, r; word divisions kept; quid = perchè) rebuilt by aligning R1101 with its clear copy 7a (Vestigia 1283) and R1104 with 13a (Vestigia 1295). R1101 and R1106 have contemporary clears (7a; 26a/26b, Vestigia 1319/4004). R1102 (16 Jul 1482), seven lines with no clear copy, read in gist: essendo in Buda per partire per venire a Posonia, sapi per gran secreto … como [X] haveva scripto ad Francesco Quirino … voleva havere seco bona a[micit]ia, et voleva dargli Vegla, et voleva lo figlio (e bastardo) per loco capitaneo de la [arma]ta per mare, et gli voleva dare ogni anno cento milia: the Venetian side's offer to Matthias (Veglia, a fleet command for János Corvin, 100,000 a year). Open: the nomenclator group [X], the informant's title, a dozen signs in ll. 3–4; R1103's faded four-line block gives fragments only. The twelve 1483 letters (b. 1/10, R1107–R1118) all have contemporary decipherments (filed copies 4a/4b, 9a, 20a, 21a, 24a, or clear slips pasted on the leaf of R1109, R1116, R1118): nothing unread there	targets/sadoleto1482/ · write-up
Henri IV → Béthune (Rome), 9, 10 and 22 Nov 1601, BnF fr. 3484 nos. 7, 8, 12 (catalogue item 22, class A)	1601	17 Sept 2026	partial	10 Nov deciphered, key largely recovered (17 Sept 2026): the notice's "copie du n° précédent" (f. 36) is the minute of the 10 Nov letter in clear; EM alignment of cipher and minute, with the marginal decipherments of 8 Nov and 10 Dec, gives a Villeroy-office key of the design Bazeries printed for Béthune's brother in 1599 (letters with homophones, dotted/comma letters for words, two-digit figures for names: 48 Cardinal, 17 Aldobrandin, 7 le Roy d'Espagne). Ten thousand écus to Cardinal Aldobrandini for the Peace of Lyon and ten thousand more promised, "sans avoir esgard à l'arrest obtenu par lad. duchesse de Nemours". The 9 Nov letter (one page in cipher) reads about six words in ten after a 4× re-transcription that fixed the doubling sign φ and the syllable groups 61 qui, 63 re, 65 si, 68 tout, 71 tion; 22 Nov open. Next: the glosses of ff. 53, 56–57 for the two-digit names	targets/bethune/ · write-up
Castelcicala → Marchese di Circello and Cavaliere de' Medici, London/Paris 1816–1823, ASNa Esteri 2337 (DECODE R9553–R9589; catalogue 166 and 167, C)	1 Apr 1816 – 9 Oct 1823	21 Sept 2026	partial	One syllabic, homophonic two-part code (~2,450 groups, local alphabetical runs, 5-figure nulls) for 24 despatches, all transcribed (9,428 groups). "Marchese di Ciscello" = the recipient Circello misread. Key of 162 values (90 G, 18 X, 54 I) rebuilt from contemporary interlinear decipherments on R9566 p3, R9569, R9571, R9589; 45% of groups valued, no letter reads through. R9586/R9588 another code	targets/castelcicala1816/ · write-up
Łaski, King John's side and a Bavarian at Wardein → the Bavarian court, Łask/Hungary/Wardein 1531–1537, BayHStA Kurbayern Äußeres Archiv 4591 ff.226–228, 257–261 (DECODE R9403, R9414, R9415; catalogue 162)	17 Sep 1531; 18 Apr 1533; 14 Mar 1537	22 Sept 2026	partial	R9403 Łaski to a ducal councillor, read 99.5% with his System B key (e/r swapped, DAG = King John); R9415 King John's side to the dukes via Augurelio, homophonic key rebuilt from its gloss, 95.8% of words; R9414 Wardein to Duke Ludwig, German homophonic key rebuilt, 92.2% after a fourth image pass (5 Oct 2026); all three 1,760/1,860 (94.6%). Open: the R9414 word signs Ψ and barred ∂ (no key), ~55 words that fail on the signs, 17 scattered R9415 words	targets/kaa4591b/ · write-up
Henry Brooke, Lord Cobham (Ostend) → [Walsingham], 20 and 22 Mar 1587/8, with a Middelburg report, BL Harley MS 287 ff. 70–72 (DECODE R8482–R8487, catalogued as "unknown sender to unknown recipient"; catalogue item 132, class A)	1588	21 Sept 2026	partial	(21 Sept 2026; reclassified 22 Sept 2026: no measured share of the text read). The "unknown" letters are signed H. Cobham, peace commissioner at Ostend; no decipherment on the leaves, none in CSP Foreign. The graphic-sign cipher is a simple substitution with a few homophones, broken from the clear words around the runs; every run reads in sense (Parma's new canal Ghent–Eeklo–Sluys, the garrison "too weake by the half", the prisoners Pigot and Barney, "a means to conclude an irreligious peace"); nomenclator signs 15 = Parma, □ = the States; 42 unidentified, .7. not firmly identified, the Jesuit's surname on f. 71r and one verb on f. 70r open. f. 11 (R8477), a different code of about eight groups, not read.	targets/harley287/ · write-up
Ferdinand the Catholic → Jerónimo de Vich, AHN Estado 8715 N.45 (Seville, 4 Apr 1511), N.57 (Burgos, 5 June 1512), N.60 (Logroño, 1 Sept 1512); N.73 (Pedrezuela, 26 Oct 1515)	1511–12	18 Sept 2026	partial	Read in part, 3 of 4 letters in substance (18 Sept 2026): a nomenclator (homophonic alphabet of figures and marked letters, plus three-letter code groups for words and syllables: pef que, mix papa, sap venecianos, fuq Francia), rebuilt by aligning the deciphered siblings N.46 and N.52BIS with their clerk decipherments, then used on the three cipher-only letters; 153 groups recovered, roughly 3–8 % of groups unread (5–14 % counting garbled stretches). N.45: Julius II's new cardinals, the Venetians absolved "sin dezirme ni comunicarme", draft Emperor–Venice terms, Ferrara, Fabrizio Colonna. N.57: how the army was pushed into Ravenna against orders, "echar la soga tras el caldero", Ferrara to be treated "sin que la sienta el papa". N.60: Massimiliano Sforza for Milan, the papal claims to Parma, Piacenza, Modena and Reggio resisted, "las armas espirituales" against Louis XII. N.73 is a different system, not read. Terrateig 1963 (vol. II prints Ferdinand's letters to Vich from this legajo) not consulted and may already print these	targets/vich1511/ · write-up
Mondoucet → Charles IX, Brussels 13 July 1572 (BnF fr. 16127 ff. 60–61), and the 1573 letters of 4 Jan (Antwerp, ff. 126v–127v) and 9 Sept (Amsterdam, ff. 135–137) (Gallica sweep find, on no list; catalogue items 1–2)	1572–73	17 Sept 2026	partial	13 July 1572: the aligners' failure traced to a glyph segmenter that split letters (1.3–1.4 glyphs per letter); by hand one glyph is one letter, key built from the glossed 16 July crib (ff. 62–64), French 5-gram beam decoder reproduces the decipherer's gloss, but the despatch itself reads only as scattered words (1–8 %, re-checked 21 Sept 2026 with the α split; the key, fitted on one 863-letter crib, is the limit). 9 Sept 1573: verbatim Court decipherment ff. 139–141, aligned glyph for letter (key1573). 4 Jan 1573, no decipherment in the volume: decoded ciphertext-only with the 1573 key (control 79 % letters after splitting the open alpha = r from round a), ~60 % read; then found printed in clear in Didier, Lettres et négociations de Mondoucet i (1891) 141–142, which confirms and completes it (21 Sept 2026)	research/gallica_sweep/mondoucet/ · write-up
Comte d'Affry (French ambassador, The Hague) → Rouillé and Bernis, eleven intercepted code letters, 6 Jan 1757 – 17 Jan 1758, KHA The Hague, Prins Willem V inv. 192 (DECODE R1054, R1064, R1065, R1067, R1068, R1070–R1074, R1076, R2067, "non-decrypted"; catalogue items 211 and 213, class C)	1757–58	19 Sept 2026	partial	(19 Sept 2026; 86% of groups measured, below the read bar). The siblings' "solutions on separate sheets" are among the DECODE images after all: Lyonet's clear decipherments of eight letters (R1052, R1053, R1062, R1063, R1065, R1066, R1069, R1075; R1065 was listed as unsolved) were transcribed from the scans and aligned group by group, giving a 651-group key of a two-part word-and-syllable code (~1,200 groups, homophones) used for the whole year. Nine undeciphered letters read at 60–95% (87% of their groups keyed); R2067 transcribed from the images. Bussemaker's 1906 extracts (19, 22 Jul, 1 Apr 1757) recur group for group; Coquelle (1904) prints the Gouvernante's words found in R1067. DECODE digit slips corrected (looped 8 read as 0). Open: R1071 (to Stainville, another code), ~5–12% of groups per letter, R1072 weakest.	targets/affry1757/ · write-up
Champagne news-letters → the duc de Nevers, Paris and Champagne, BnF fr. 3623 nos. 23, 24, 25, 60, 78 and fr. 3625 no. 55 (DECODE R9453–R9455, R9459, R9460; catalogue item 30, class A)	Feb 1590 – July 1593	21 Sept 2026	partial	No. 23 (Bienvenut, 28 Feb 1590) is a clear copy with a nine-entry name code whose key is bound on f. 36r: read in full. Nos. 24, 25, 60 (June 1590) share a two-digit homophonic alphabet laid out in alphabetical order, rebuilt from a few interlinear glosses: 10 of 12 runs read (la rupture de l'edict [de] l'union, contre la ligue, du sr Ludovic). Tomokiyo's key no. 46 does not fit. Open: two runs of no. 24, code numbers, no. 78's one group, Laurière's 9 Jul 1593 letter (f. 66r) deciphered in full on 29 Sept 2026 with his own key sheet, Tomokiyo's Nevers key no. 57 (fr. 3995 ff. 102–103), identified by setsunaatto in issue 13 (vous estes necessaire pres du Roy afin qu'il ne retarde d'ung jour son assemblée); our first-pass table had been one group out of step. The key also reproduces the office's interlinear decipherment of his 13 Jul letter (no. 10) in a blind test contributed by setsunaatto (PR 15: 80% of the gloss in matching runs of 4+ letters, shuffled controls at most 30%), and reads the one run the office left unglossed. On 2 Oct 2026 key no. 57 deciphered La Verrière's letter from Poissy of 12 Aug 1593 (fr. 3985 f. 58, the letter Tomokiyo names for the key, glossed at the time): 222 of 248 cipher tokens (89.5%), on the prince de Conty, Madame d'Angoulême and the Montmorency house, La Trémoille, and a papal design against the Cardinal de Bourbon; 26 tokens open at that pass. 3 Oct 2026: f. 58 re-read from lossless regions at Gallica's maximum, the key's name panel f. 103v read in full (Conty, Montmorency, Cardinal de Bourbon, Lorraine all on it, now grade H), the writer's caret = p (port[ée]), 29 barred = the Spaniards, 252 = lettre: 235 of 248 (94.8%), run A 98.2%; 13 open (6 illegible signs, 7 with certain key values and no sense, 340 tabour among them; Cotgrave/Nicot checked; 348 voir a conjecture, not counted). Nos. 24/25/60: key no. 57 does not fit them; a bracketed exhaustive search of their five unknown letter values verifies none on a second occurrence (63 of 104, unchanged). All counted letters: 383 of 444 cipher tokens (86.3%), partial.	targets/champagne1590/ · write-up
Lord Cobham (Ostend, Bourbourg) to Walsingham, BL Harley MS 287 ff. 80–97 (DECODE R8490, R8492, R8495, R8496), catalogue item "Cobham (Belgia) to Walsingham", class A	5 May – 9 June 1588	21 Sept 2026	partial	Key rebuilt from the interlinear glosses of Cobham's April letters (R8488, R8489, R8493) as a sign inventory; the catalogue's R8497 is Bodley's 1590 cipher. Code numbers glossed (15 = Parma). A candidate-set solver over sign transcriptions reads about a third of the cipher words: "in all armadas", "so forwarde in the action", "the haven of Rain", "Italians arriving … Antonio", "within France".	targets/cobham1588/ · write-up
Riksarkivet group, 14 DECODE records R4282–R4341 (catalogue 207, class A): archbishop Johann Friedrich of Bremen → Johan Adler Salvius at Hamburg (R4330, R4331); key-box ciphertexts Chifferklaver låda II:113, 114, 133 (R4282, R4284, R4306); Rusdorff → Oxenstierna (R4333–R4337); Oxenstiernska saml. E 708 (R4338–R4341)	1628–1633	21 Sept 2026	partial	The Bremen letters read in full; three ciphers attempted, open. Catalogued as one Latin cipher of an unknown sender, 1628–1646; in fact four unrelated groups, and E 708 is June 1633, not 1646. Bremen, 27 Nov and 2 Dec 1631, to Salvius (not Åke Tott): two-digit homophones, four consecutive numbers per letter, the bands taken down the columns of a six-column alphabet grid (8 a, 12 g, 16 n, 20 t, 24 b … 96 s, 100 z), rebuilt from the clerk's interlinear glosses (≈80% of groups); k, z, m, w, y fixed from unglossed words and q, x predicted by the grid; every group reads. The archbishop begs for succour against Gronsfeld's and Bönninghausen's troops. Open: R4306 (to Amsterdam 1632, IoC 0.078, no language fits under annealing in ten languages), R4284 (Latin, 108 numbers), R4282 (Latin letter cipher, ~33 signs); two box keys (R4296, R4310) tested and rejected. Rusdorff (5) and E 708 (4, glossed at the time) not worked	targets/riksarkivet1628/ · write-up
Duke of Sessa (Luis Fernández de Córdoba) → Charles V, Rome, 18 Apr 1524, and its duplicate, RAH Salazar A-31 (9/31) ff. 128–131 (DECODE R9877–R9878, catalogued "Luis Fernández, April 1424, Non-decrypted")	1524	18 Sept 2026	partial	Misdated in DECODE by a century, and read in part. The volume is Bergenroth's Salazar A. 31 (CSP Spain vol. 2); the letter is dated on f. 128v and Bergenroth skipped it. Eight sibling despatches in the same cipher carry the court's decipherment (1523–25); about 110 code groups and the letter alphabet were rebuilt from them by token tiling and template matching (rad dicha, lif mostrar, z y -on, hud plática). The nine cipher lines read at the word level: las dichas […]s no le las mostraron, que le ponen sospecha que desea aver algo … hablé en la ora a Su Santidad a pedirle que haya provisión … para los casos de suyços … pues tiene color para ello, aviendo començado la plática de la paz y tregua. 2 Oct 2026 (pass 16): two more Sessa despatches transcribed in full and deciphered in part with the same key — R9873 (14 Apr 1524, A-31 ff. 79–86; CSP ii 636–638 used as a topic check) 2,165/2,473 tokens read as sense (87.5%), and R9898 (24 Feb 1525, the day of Pavia, A-34 ff. 150–156) 2,361/3,059 (77.2%); 3 Oct 2026, pass 17: open groups pooled across R9873, R9898 and R9660, tentative values (cas tiene, var agora, rac dize …): R9873 2,295/2,474 (92.8%; 87.8% strict), R9898 2,441/3,059 (79.8%; 77.9% strict), still partial; R9898 is not without decipherment, as first noted: six cipher runs carry contemporary marginal decipherments, which gave v em cierto, li mi, net halla, sud datario and Micer Agostino Foyeta. They also fill f. 128's residue: de lo susodicho, aunque no (rus dicho, hay aunque tentative). Open: vo, per and one spelled clause on f. 128; about 260 single-context code groups in R9873/R9898 and R9898's bleed-through pages	targets/sessa1524/ · write-up
Bavarian key volume, fourteen ciphertexts → the Dukes of Bavaria, Buda/Kraków/Rome/Pressburg 1529–1583, BayHStA Kurbayern Äußeres Archiv 4591 (DECODE R9291–R9427; catalogue 161, B)	24 Nov 1529 – 12 Jun 1583	21 Sept 2026	partial	R9416 f. 263, broken ciphertext-only, is split out under ciphertext-only. Separate letters in ≥6 systems. Read: R9325 (Fulda protest 1576, key R9324 in the volume); Łaski/King John R9322, R9417, most of R9323 (key rebuilt from glosses); R9416 f.262 read at the time. Broken, read in part: R9416 f.263 (Pressburg talks, ciphertext-only with new de-1500s model); System A′ R9408 (Jörg Weinmeister, Ofen 8 Feb 1534), R9409, R9410 f.244r (1 Apr 1535) and ff.246–247 (King John Zápolya to the dukes, Latin, Buda 30 June 1534, transcribed and deciphered 2 Oct 2026), R9413, R9427: key first from R9427's gloss ('ewr F.G.'), then the R9407 key (augurelio1535) carried back and re-learned per letter, 89.3% of 6,827 words with sense (5 Oct 2026; 84–94% per letter: R9413 94.3, R9408 92.0, R9410 Latin 88.9, R9409 87.2, R9410 f.244r 86.1, R9427 84.3; K word sign = King Ferdinand by context, grade B; new names Bebek, Debreczen); R9367 (22 Mar 1535, key register R9423 block 2_4). Open: R9424 (1583, attempted); R9319 dot grille blocked (cover letters f.115–118 not imaged); R9291 a key fragment	targets/kaa4591/ · write-up
King John Zápolya and an agent at Buda → the dukes of Bavaria (filed on DECODE under Cornelio Sperantio, the addressee of the cover), Buda 1534, BayHStA Kurbayern Äußeres Archiv 4591 ff.248, 250–251 (DECODE R9411, R9412; catalogue 163)	6 Feb 1534; 8 Feb 1534	22 Sept 2026	complete	R9411 King John to the dukes, Latin, on the Hungarian gold and silver mines and his servant Nicolaus Freiberger: key rebuilt from its interlinear decipherment of the time, 340/356 cipher words (95.5%). R9412 German newsletter on Werbőczy's everlasting peace with the Turk, Peterwardein, Titel, Salankamen, the mines: same key, 546/573 (95.3%). Together 886/929 (95.4%, measure.py, 5 Oct 2026, cipher words only, supplies counted unread). The key is System L of R9415 (kaa4591b): one correspondence. Open, illegible: 16 words in R9411 (13 short gaps at blotted signs); 27 in R9412 (the signature and several names among them)	targets/sperantio1534/ · write-up
Andreae, Chymische Hochzeit Christiani Rosencreutz, Strassburg 1616, pp. 73, 95, 97, 116, Embassy of the Free Mind (Ritman Library) copy (DECODE R4502, catalogued as 'Johann Valentin Andreae to unknown recipient'; catalogue 210, class C)	1616 (printed)	22 Sept 2026	complete	Solved earlier by The novel itself and Foxcroft 1690. Not a letter: the four 'cipher pages' are the printed inscriptions. The two long ones are a simple substitution in an invented 19-sign alphabet, glossed by the page boy on pp. 96 and 98 ("hie ligt begraben Venus…", "wann … verschmelzen … Mutter eines Königs"); sign table rebuilt here and decodes all 156 letters. The two sign lines are dates, 1378 and 1459, per the crcsite.org edition of Foxcroft (graded C sign by sign)..	targets/andreae1616/ · write-up
The Cardinal-Infante Ferdinand (Brussels) → Ferdinand, King of Hungary, AGR Secrétairerie d'État allemande 540 (DECODE R1889, "Non-decrypted"; draft R954)	16 Nov 1635	9 Oct 2026	complete	Split entry, part 1 of 3; independent re-solution (Aymeloglu, Sept 2026). The Cardinal-Infante's own draft (DECODE R954, same shelfmark and date) gives the passages to be enciphered underlined; aligning its 74 words with the letter's 463 tokens gives a Latin alphabet of two to five numbers and letters per letter. 450 of 463 tokens (97.2%) give the draft's letter, the rest copy slips. He refuses imperial winter quarters in the archbishopric and city of Trier, under Spanish protection for two hundred years, and keeps the Spanish garrison there	targets/ferdinand3/ · write-up
Ferdinand, King of Hungary (Stuttgart) → the Cardinal-Infante Ferdinand, AGR Secrétairerie d'État allemande 540 (DECODE R1887, "Non-decrypted", no transcription)	28 Oct 1634	9 Oct 2026	partial	Split entry, part 3 of 3; independent re-solution (Aymeloglu, 23 Sept 2026, also partial). Transcribed here (474 tokens): a different repertoire of vowel-series letter groups, numbers and graphic signs. The last cipher passage repeats the clear words after it (promovere haud dedignetur), which fixed five families; a crib-seeded anneal and context gave the rest. 88.4% of tokens read as sense (93.9% with conjectures), below the read bar: after Nördlingen, the King asks his cousin to draw the enemy apart by a diversion of his arms. Open: a 6-token word after Eventus vero inde and a 16-token span before negotium (every sign valued, no Latin fits, last dictionary + LM pass 9 Oct); conjectures ausus quos, eventus, partibus, quam ea	targets/ferdinand3/ · write-up

Read after matching with key from external sources

An existing key not previously tied to the document was found to fit and applied.

Target	Date	Landed	Extent	Result	Where
Lodovico Birago to Duke of Nevers, 5 April 1571 (Français 3252 — ff36r–37r)	5 April 1571	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 3/3 pages transcribed; 3/3 cipher pages deciphered in part. 923 elements transcribed, 805 identified (87.2%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/birago1571/ · write-up
Ercole Rondinelli to Duke of Nevers, 11 November 1590 (Français 3613 — Rondinelli, ff62r–v)	11 November 1590	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/2 pages transcribed; Deciphered. 656 elements, 655 identified (99.8%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/rondinelli1590/ · write-up
Duke of Nevers to M. de Gondi, 27 March 1594 (Français 3622)	27 March 1594	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 3/3 pages transcribed in part; Deciphered. 245 elements, 230 identified (93.9%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/nevers1594gondi/ · write-up
The Doge and Senate of Venice → Giovanni Mocenigo, ambassador with Henry III (Doge Pasquale Cicogna; subscribed by the secretary Marco Ottobon), BNE Mss/994 ff. 34–38 (DECODE R2252, "N/A", cleartext language given as Latin; Tomokiyo's list, Valle de la Cerda's lost solution)	27 Apr 1589	9 Oct 2026	partial	Deciphered in part with the Council of Ten's Zifra Prima of 1577–78 (ASVe CX Cifre b. 4 reg. 16 ff. 64–65, DECODE R1789), plus two homophones found from context (d83 = a, c86 = o): 981 of 1,565 cipher groups (62.7%, measured). The first-pass transcription (130-dpi scan) misread 7 as 2, 8 as 0 and confused 3/5/8; corrected by hand (721 groups, 46.0%), then by a noisy-channel decoder calibrated on the hand reading (89.3% agreement on held-out pages) and by fitting proposed lines, with a wrong-line control (10% of groups fit by chance vs 80% on the true lines). Two pieces: the Senate to Mocenigo on the French ambassador's request that Venice help Henry III and move the Emperor to aid him; and the copy of the Senate's answer to the French ambassador (de Maisse): Venice will intercede with Sixtus V to quiet the risings and "ammorzar l'incendio di guerra". Open: ff. 35v–36v (29–48% read) and gaps on ff. 37r–38r, digits the scan does not separate; needs a 400-dpi BNE scan, the Senate register (ASVe Deliberazioni Secreta 1589) or de Maisse's letter-book (BL Add MS 30635–37)	targets/ottobon/ · write-up
Unidentified writer (Hamburg) → the Hessian chancellor (Kassel), despatch with code numbers and two letter-cipher passages, HStAM 4 f Dänemark Nr. 125 ff. 2–4 (HCPortal record 494, nomenclator, "partially solved"; catalogue 341, B)	4/14 May 1672	9 Oct 2026	partial	Both cipher passages deciphered with the letter table of HCPortal key 255 (HStAM 4 d Nr. 1234, Secretary Lincker 1666/1676), found by screening the 106 Marburg keys; 59 of 64 cipher tokens (92.2%). The 26 code numbers carry a second hand's glosses (601 King of Denmark, 651 Brandenburg, 768 Sweden, 634 Holstein, 303 alliance; 602 Braunschweig inferred); the key's own word list does not fit. Passage 1 "Totaliter disgustirt und … in Holstein", passage 2 "Gott geb das all wol ablauff". Open: five signs (e r t a z) that make no word; the glosser got the same letters and the 1672 key (which differs from the 1666 one in at least one cell) is not available.	targets/hesse1672/ · write-up
London → Johann Georg von Ilten: unsigned letter of 15/26 Feb 1743 and Ernst von Steinberg, 29 Mar / 9 Apr 1743, GWLB Hannover Ms XXIII 1234:29,2 pp. 73–75 and 119–121 (sibling of catalogue 346)	Feb–Apr 1743	24 Sept 2026	complete	Deciphered with the Ilten papers' own code table, 220 of 223 groups. The unsigned letter's 189 groups had no decipherment. The "Chiffre-Tabelle" in Ms XXIII 1234:31,1 pp. 386–391 (863 numbered groups, blanks null) matches the office worksheet of Steinberg's letter, 28 values, and reads both: the King approves the forage instructions to Sommerfeldt and wants the forage eaten so the French find none; an extract of Lord Stair's instructions. One group lost in the binding.	targets/ilten1743/ · write-up
Johan Isaac de Swart (St Petersburg) → Pieter van Bleiswijk, Nationaal Archief 3.01.25 inv. 610 (DECODE R1036), with the letter to the States-General of 29 Nov 1787, NA 1.01.02 inv. 7408 (DECODE R1040) (catalogue 224, class 2.7)	8 Mar 1782	21 Sept 2026	complete	Deciphered, 96.7% of 1,719 groups. The "unidentified" codebook is Croiset's 1765 key for the Russia legation (DECODE R1038), parsed with its digit marks to 5,961 entries; 34 values recovered from context. The French and Spanish answers to the Austro-Russian mediation, Catherine II's Greek project, Prussia's accession to the Armed Neutrality. 106 rare groups open. R1040 read in part: its second part reads with the same key, while the first 1,928 groups are in a missing codebook	targets/deswart1782/ · write-up
András Bay → Ferenc Rákóczi II, Jászvásár (Iaşi), MNL OL G15 Caps. D. Fasc. 81. 5 (DECODE R478; catalogue 51, class C)	8 Mar 1706	21 Sept 2026	complete	Read with Rákóczi's key DECODE R609, unchanged. Found by scoring all 94 G15 key transcriptions against the letter; DECODE's transcription corrected in seven places from the images. 414 groups, 413 in key, 50 = ba from context; ~10 groups on p. 2 without clear sense (97.6% read as sense). Muscovite and Polish envoys at the Moldavian court, the Tsar suing Charles XII for peace, Pekri's message to the voivode, the Swede at Vilna	targets/bay1706/ · write-up
A correspondent at Randan (Auvergne; probably the comte de Randan) → Louis de Gonzague, duc de Nevers, BnF fr. 3975 ff. 28, 30 (DECODE R4285, R4286, catalogued as "Unknown sender (Rome)"; catalogue item 185, class C)	27 Mar – 4 Apr 1587	22 Sept 2026	complete	(26 of 27 cipher units accounted for, 0.963, measured). Two clear letters with seven cipher items. Key: Nevers key book BnF fr. 3995 no. 11 (f. 23), unchanged: 94 = the marquis, 6 = M. d'Entragues, 23 = the queen of Navarre, 45 = the duc d'Épernon (glossed). The run and the dateline read only after the writer's glyph forms were learnt from his glossed letter from Clermont, BnF fr. 3413 f. 102 (Feb. [1586]; Tomokiyo gives it to the cardinal de Guise): Ŧ = p, capital F = et, 1 = d. The run reads "passant pres Lyon et ce pais" (Épernon's alarm); the first letter is dated in cipher "De Randan", not Rome; two signs after it are a struck-out false start. Open: one sign taken as a null	targets/nevers1587/ · write-up
Charles I (Oxford) → Prince Rupert, BL Add MS 18983 f. 14 (DECODE R4921; catalogue 71, class C)	29 Apr 1645	21 Sept 2026	complete	(21 Sept 2026). Holograph, marked Duplicate; DECODE "Non-decrypted", no decipherment in print (Warburton 1849 prints only Digby's letters of 27 and 30 April). Read with George Lasry's key TNA SP106-5 (DECODE R929, the King–Queen cipher) applied unchanged: the King cannot convoy his artillery train and asks Rupert to march and fetch it. Groups 81–98 outside the key (95 = my, 98 = artillery from context); about 15 code groups open	targets/rupert1645/ · write-up
Bordeaux → Brienne, London, 30 May 1653 NS (Thurloe intercept, BL Add MS 4200 ff. 88-90; 813 tokens as corrected, 810 and 156 symbols in Tomokiyo's transcription)	1653	18 Sept 2026	complete	(18 Sept 2026) with the English Deciphering Branch's key sheet "To Mr. Bordeaux French M.r at London 1653, Key & Calculation" (BL Add MS 32263 f. 1, DECODE R7537): plain 15-37 = a-z, and the prime, two-dot and overbar series are alphabetical syllabaries with words filed in place, the design the 16 Sept solver assumed. The ciphertext-only failure is explained by the transcription: checked on the R8390 images it merged two signs (ꝺ = u, ∂ = n), dropped tokens and marks (21 corrections, one from the duplicate R8393), and the clerk often omits the mark himself. The Council heard the Bordeaux (Ormée) deputies through two commissioners before Cromwell ("le General") gave the envoy his audience; the treaty is unsigned; England may help Bordeaux indirectly, so hold back French mediation. Open: six nomenclature codes (probably Stuart names among them) and a dozen rare tokens	targets/bordeaux/ · write-up
"Ferenc Rákóczi II to unknown recipient", four letters, MNL OL G15 Caps. C. Fasc. 39 and P237 Festetics 10. d. 2. 42–44 (DECODE R483, R852, R912, R922; catalogue 44, class B)	1706–1711	21 Sept 2026	complete	Read; three of the four were written to Rákóczi. Second pass from the 14 page images: every cipher page re-transcribed, and the key sheet DECODE R639 read (45 = S, 47 = T, Z line, unnumbered 173/174 = de). R912 and R922 (Poland and Danzig, Oct–Nov 1707: Stanisław, Charles XII, the Tsar, Rehnskiöld, the Palatines of Belz and Ruthenia, Tököly) keyed at 2,290 of 2,352 groups (97.4%); about 820 of them, chiefly R912 p. 1, already had a contemporary interlinear decipherment. R852 (Danzig, 22 Jan 1706) has its own clear copy and an earlier issue of the table. R483, Rákóczi’s Latin letter signed at Lwów on 27 June 1711 against the Peace of Szatmár, was deciphered interlinearly at the time; its 224-code syllabic key rebuilt here by aligning the cipher with that decipherment (70% letter agreement).	targets/rakoczi1704/ · write-up
Dr Nicholas Wotton (Soissons) → Queen Mary I, BL Harley MS 1582 ff. 8r–10v (DECODE R8499, catalogued as "N. Wottoy ? to unknown recipient"; catalogue 120, class C)	13 June 1554	22 Sept 2026	partial	(75.0% of 2,801 non-null signs firm, 89.4% with tentative; 68.9% measured as sense on v9, 5 Oct 2026). Key found on DECODE as R354 (TNA SP 106/2 f. 162, "France 1554"): homophonic sign alphabet, nulls, ~140-entry nomenclator; eight transcription and key passes. Carew and the rebels at Compiègne with the French king and the Constable, raised pensions, a band with wages, the Isle of Wight and Plymouth, a repentant rebel met secretly in Paris. Open: the rest waits on the contemporary deciphered copy at ff. 5r–7v (not digitised; BL order) and ~40 rare code signs absent from R354	targets/harley1582r8499/ · write-up
Robert Reade (Paris) → his cousin [Thomas Windebank?], 19 Apr 1641, BL Harley MS 7001 ff. 148–149 (DECODE R7766; catalogue 136, C)	1641	22 Sept 2026	partial	(93% read, 7% conjectured). Writer from signature and endorsement ("my Cousin Reade from Paris"): Robert Reade, Windebank's nephew and secretary. Key: the Deciphering Branch reconstruction of the Windebank cipher, Add MS 32256 f. 4 (DECODE R9115). The two passages are on Windebank's return under a summons and on St Ravy and Mr Forster. 193/215 groups read from key values (562 = strong, from the key sheet), 8 from rebuilt values; 14 code groups blank on the key are conjectured	targets/harley7001/ · write-up
Prince Frederick of Orange (London) → the Hereditary Prince, 7 May 1795, KHA The Hague, Koning Willem I inv. XVIII-3 (DECODE R2242, "non-decrypted"; catalogue 212, class C)	7 May 1795	21 Sept 2026	complete	Deciphered with the R1892 key, unchanged (21 Sept 2026). Same vertical digit-pair square; page 1's clear Dutch under the cipher confirms it and fixes 12 word signs. Pages 3–4 (late May or June 1795): scattered émigré troops, the ratification of the Treaty of The Hague with France published before it arrived, from letters of the ministers Meyer and Blauw in Paris, the envoys De Sitter and Van Grasveld sent to carry it (identified in Colenbrander, Gedenkstukken I pp. 655, 678; the letter itself is not printed there), "Robespierismus". 2 Oct 2026: p3 l1 and p4 l8 re-transcribed at full resolution, 12 more word signs from context (dat, daar, naar, zonder, Engeland …): 981/1,037 tokens read as sense (94.6%), 97.9% given a value. 3 Oct 2026: one pair corrected (saamenhang); stricter measure (context-only sign values count unread; y scored as ij; a word also passes if attested in period Dutch, Colenbrander's Gedenkstukken; shuffled controls 8.4%), seven sign values upgraded where the sentence admits one word: 994/1,037 read as sense (95.9%), over the read bar. Open: 20 single-occurrence word signs and four context-only values (no key sheet, not in R1892, not in print), two ink-blotted tokens, "Dankaert" in the P.S. unidentified.	targets/r2242/ · write-up
Philippe Groffey (?) → Ferenc Rákóczi II under the alias Pompeio Cesoni, MNL OL G15 Caps. C. Fasc. 39 pp. 277–279 (DECODE R902; catalogue 48, class C)	15 Oct 1707	20 Sept 2026	complete	Deciphered with the archive key R639; the catalogue direction reversed. The clear address is A Monsieur / Monsieur Pompeio Cesoni, making Rákóczi the recipient. The matching key is DECODE R639, Fasc. 44/08, headed De Monsieur de Bonac et Graffei: a 10–560 French homophonic syllabary and nomenclator. It identifies 1,020 of 1,108 transcribed groups (92.1%), after validation on interlinear siblings R852 and R912. The report concerns Ráday's letters to Rehnskiöld, work at the Swedish and Polish courts, Prince Konstanty, and Polish liberty oppressed in favour of King Stanisław. The key heading, subject and diplomatic role make Philippe Groffey the probable unsigned writer. Both cipher pages were then transcribed afresh from the images, group by group (30 Sept, 5 Oct 2026): 982 of 1,017 groups (96.6%) read as sense (measure_tr.py), twelve after a one-figure correction forced by context. Open: ten groups ending the first paragraph that make no word, about 25 scattered single groups; the attribution remains probable, not certain. Manuscript images are not reproduced because archive permission is required	targets/rakoczi1707/ · write-up
Duke of Sessa (Luis Fernández de Córdoba) → Charles V, Sessa and Rome, seven despatches, RAH Salazar A-27 (9/27) ff. 141–441 (DECODE R9660–R9666, catalogued "Luis Fernández? (Sesa) to unknown recipient"; catalogue 146, class B)	20 Feb–27 Apr 1523	21 Sept 2026	partial	Solved earlier by Bergenroth 1866 (CSP Spain vol. 2 nos. 534, 540, 541, 544, 545), from the court's contemporary decipherments in the same volume. Mostly in print; the rest read in part. Four records are those letters or their duplicates and one (545) is clear. The cipher is the 1524 Sessa nomenclator and its key applies unchanged. Open: R9660 (20 Feb) and R9665 (26 Apr) are uncalendared, with one paragraph of R9660 read as a sample.	targets/sessa1523/ · write-up
Ormanetto / Clementino, Spain nunciature to the Secretariat, ASV Segr. Stato Spagna 10 (DECODE R117–R118; catalogue #239, class A)	15 Oct 1576 or 1577	1 Oct 2026	complete	R118 read with the same key (Devadas, 1 Oct 2026): the nuncio has told Secretary Pérez what Rome wrote in the cipher of 19 September. The clear note on the leaf is the contemporary decipherment, minus the dateline; our 21 Sept crib failed because it was misread. R117 is a clear memorial, no cipher.	targets/ormanetto1576/ · write-up
Philip II to the nuncio Ormanetto, copy sent to the Cardinal of Como, ASV Segr. Stato Spagna 7 ff. 303–304 (DECODE R116; catalogue #259, class C)	spring 1573 (docket 17 Jun 1573)	1 Oct 2026	complete	Deciphered with the Spain nunciature's ordinary cipher, key reconstructed by Ajaydas Devadas (email 1 Oct 2026); 1,298 of 1,324 cipher tokens (98.0%) measured as sense after a re-transcription from the full-resolution scans (5 Oct). The 21 Sept attempt failed because the cross-stroke on 2/4/5 is a word-end mark and an undotted 0 closes the digit before it. Our own transcription decodes to Italian (−3.0 per letter against a best of −6.7 for 200 shuffled keys): the King on joining the Pope's league with the Emperor, and on Pitigliano. Six single-context code groups open (32, 2̇88, 84, 42, 53, 3̇22 as per).	targets/ormanetto1573/ · write-up
Fra Guglielmo Vizani → the Comte de Césy, Prinkipo, BnF Français 16158 ff. 292–294 (Cryptiana's unsolved list, "no transcriptions")	8 Oct 1637	1 Oct 2026	complete (six name codes open)	Read with an outside key (Satoru, with Claude, 1 Oct 2026), verified here. Eight cipher lines, 54 words, a homophonic symbol cipher: 4,000 scudi promised for Monsignor Beria, "close to obtaining the Patriarchate for himself", who hopes the Patriarch will be replaced within two months. 46 of 48 plain words reproduce from his transcription; the decode scores −1.78 per letter against −4.30 for the best of 200 shuffled keys. Name codes 31, 36, 64?, 198, 227, 231 unread.	targets/vizani1637/ · write-up
"El Gilino" → Francesco II Sforza, Duke of Milan (at Cremona), Paredes de Nava, BL Cotton MS Vespasian C IV ff. 214–218 (DECODE R8572, "Non-decrypted", place "Daredos"; tracker item 13)	15 Sep 1527	9 Oct 2026	complete	Barred G/B/D/o signs with two-digit syllable numbers; a scan of the Sforza key sheets on DECODE found the key, DECODE R5258 (ASMi Carteggio Sforzesco cart. 1591 no. 24, "Cum equite Bilie"), filed under another agent. With a few context values 2,872 of 2,989 signs (96.1%) give Italian: Gattinara out of favour over intercepted letters about Genoa, Bourbon's death, the Pope's liberation said to be revoked, Andrea da Burgo pro opprimenda Italia, the investiture of Milan sent to the Infante with 100,000 ducats. George Lasry had independently rebuilt the same key from the ciphertext (7 Oct 2026; agrees with R5258 on 101 of 106 numbered signs); his nine-page edition and full reading are on the page. Open: 9 signs and 45 doubtful words	targets/gilino1527/ · write-up

Read with known key

The key was already identified for the document or its correspondence and was applied.

Target	Date	Landed	Extent	Result	Where
[François ?] d'Entragues to Comte de Saint-Aignan, auprès du duc d’Anjou, 1581–1582 (catalogue) (Français 15906 — ff759r–760v, ff780r–781v)	1581–1582 (catalogue)	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 1/8 pages transcribed; 7 more in part; 5/7 cipher pages deciphered; 2 more in part. 87 elements, 78 identified (89.7%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/entragues1581/ · write-up
Filippo Sega to Cardinal Caetano, 9 January 1591 (Français 3613 — Sega, ff148r–v)	9 January 1591	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 1/2 pages transcribed; Deciphered. 100 elements, 100 identified (100.0%): 99 plaintext codes and 1 null. Spot-checked here 13 Oct 2026: confirmed.	targets/sega1591/ · write-up
Philip II / Archduke Albert to Duke of Sessa / Philip II, 31 March 1597 / 25 August 1596 (Français 23054 — Spanish numerical correspondence)	31 March 1597 / 25 August 1596	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 3/3 pages transcribed in part; 3/3 pages deciphered in part. 379 elements, 340 identified (89.7%). Spot-checked here 13 Oct 2026: confirmed.	targets/sessa1597/ · write-up
Antonio Bianchi (XXIII); Piero Lusco (XXXI), as written in the visible subscriptions to Signor Castrino / Castrini, as written, XXIII: Venice, 3 February 1610; XXXI: Ferrara, 3 August 1610 (Italien 2061 — ff72v–73r, 95v–96r, 97v–98r)	XXIII: Venice, 3 February 1610; XXXI: Ferrara, 3 August 1610	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 6/6 pages transcribed; 6/6 pages deciphered. 256 elements, 254 identified (99.2%). Spot-checked here 13 Oct 2026: confirmed.	targets/bianchi/ · write-up
Bossuet correspondence to Includes letters to the abbé Bossuet, Selected letters from1697–1698 (NAF 28227(9))	Selected letters from1697–1698	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 3/32 pages transcribed; 28 more in part; 12/28 cipher pages deciphered; 16 more in part. Transcribed passages: 642 elements, 626 identified. Spot-checked here 13 Oct 2026: partly confirmed.	targets/bossuet/ · write-up
Graphic-sign letter, a woman to a royalist officer, BL Add MS 72438 f. 107r (DECODE R8728; found beside Cryptiana item #6, Digby-coach keys)	before Oct 1645? (undated)	9 Oct 2026	complete	Deciphered with the key filed beside it: the lower alphabet of the key sheet f. 106 (R8727) applied unchanged, plus five unlisted homophones; checked against the partial contemporary interlinear gloss. 122 of 128 cipher words (95.3%); the four-sign heading and five words (a blot, writer's slips) open. She writes to an old acquaintance who was reported killed or taken: "God knowes how ioyfull I was to heare you got safe to the King". DECODE has no reading. The Cryptiana items themselves stay unread (row under Not solved)	targets/rupert/ · write-up
Desmarets → unnamed correspondent, Marly 4 June 1710 (DECODE R10198; catalogue, B)	1710	2026-10-09	96.8 % of 471 groups (12 code types open)	Read with Tomokiyo's 1710 table, applied to the 471 groups as one continuous stream (reported in issue #28 by yuanyi-350). The first attempt (21 Sept) closed it unread because it aligned line by line; the clear text is written out of sync with the figures. The table was built from this letter, so the check is not independent. Peace terms at Geertruidenberg: "on veut la paix, on la veut sure, mais on veut encores menager l'honneur du gouvernement"	targets/desmarets1710/ · write-up
Pieter Battier → Gaspar Fagel, Madrid, NA 3.01.18 inv. 401 (109 scans; named key NA 1.10.29 inv. 1209, DECODE R2794)	1686–1688	4 Oct 2026	partial	Read with known key. Two letters pass strict sense without a known contemporary decipherment: 24 Apr 1687 (58R–59L), conservatively 230/235 (97.9%); 8 Apr 1688, 313/323 (96.9%), conservatively 312/323 (96.6%); 322 key-attested. 19 Dec 1686 is the third validation letter: four passages, 186 tokens, reproducing the contemporary list on 32R apart from spelling and gedaan/afgedaen; this strengthens the inv. 1209 key identification. 4 Dec 1687 (94.3%) and 12 Feb 1688 (92.3%) remain below the strict bar. Earlier volume estimate roughly 55%; no revised denominator for the 235-token addition. The first two validation letters give 1113/1132 comparable tokens (98.3%; 94.7% key-alone); review is software, not paleography. LLM transcriptions from Nationaal Archief scans checked against the key images; shuffle control run with the target-folder script. Decipherment by Feyseel Nur (with Claude and Codex).	targets/battier1686/ · write-up
Prince de Conti → Laigue and Noirmoutier, Paris, 5, 26 and 27 March 1649, BnF fr. 3854 nos. 41–43 (catalogue item 26)	1649	18 Sept 2026	complete	(18 Sept 2026). Nos. 41–42 were found already read (Lasry, HistoCrypt 2023, a poly-homophonic letter cipher; his paper miscites "fr. 3584"). They are re-read here from a fresh transcription, adding ne pouvant, pour insister and the B‹l›etaigne slip. No. 43, supposed to be their crib, is a different syllabic number code: four interleaved alphabets, five-syllable blocks, underline = reversed syllable, 99–250 nulls. Its clerk never glossed the first 12 lines (264 groups). The key was rebuilt from ~3,300 glossed groups; it reads the opening, Pour Monsieur de Legue …, and nine structure/context values were then confirmed by glosses elsewhere. The whole mémoire is edited: Conti to Laigue, 5 March 1649, the day after Rueil opened, promising the Archduke to break the conference the moment he enters France	targets/conti1649/ · write-up
Conde de Hernán Núñez (Stockholm) → Baltasar de Fuenmayor (Copenhagen), four letters, 5–26 Dec 1674, AGR Brussels SEG inv. 2559 (DECODE R1012–R1015; catalogue item 42, class A)	1674	21 Sept 2026	complete	(21 Sept 2026). Clear Spanish with cipher passages (1,716 groups), each deciphered in the margin at the time. Aligning the margins showed the cipher to be the Balbases 1677 syllabic nomenclator; that key reads all four unchanged, verifying the margins and recovering words DECODE could not read (four values added: 167 Suecia, 161 resolución, os para, 6+ o). Content: the Swedish–Danish treaty project, Hamburg and Brandenburg, Sweden's embassy playing for time against French demands, the senate deciding on armed aid to France. Second pass on the images (22 Sept 2026): the 118 groups first left unread were DECODE misreadings (62), code groups fixed by the margin (48), nulls (7); one illegible, 99.9% of 1,716 groups valued (99.3% without 11 uncertain); 13 = e and underlined N = V.S. corrected; DECODE omits two lines of R1012	targets/hernannunez1674/ · write-up
Galeazzo Visconti → Angelo Bolano (Alessandria) and → Francis I (pieve de Locate), BnF fr. 3045 f. 28 no. 24 and fr. 3096 f. 91r no. 47 (DECODE R4230, R4256; catalogue item 179, class C, filed as "Français 3045 (2 items)")	15 Oct 1528; 15 June 1529	22 Sept 2026	complete	(1408 of 1477 cipher signs, 95.3%, measured; to the King 667/699 = 95.4%, to Bolano 741/778 = 95.2% after a push pass; upgraded from read in part 22 Sept 2026. Thin margin: without two judgement calls, a doubled-R loro and a dotted p read U from sense, the Bolano letter is 94.6%). Lasry's 2023 key extended: a bare Λ doubles the next letter, far = CHE or NTO, ρ = F, θ = D and G, q = H (F to the King). Both clear closes read "ali XV", not 12 as in DECODE and Tomokiyo. The Venetians fear the peace; Visconti urged forcing Milan ("il tuto, Sire, è a uno fine non bono"); Savona relieved, "come batale perduta, ogniuno voleva andare ale stancie", infantry gone for want of pay. Open: codes 48 (probably Saint-Pol), rr, qq, 42, ep (probably the King), no key; one-off stretches in fr. 3045 l.15, l.18 (a place name) and fr. 3096 L12, L18	targets/visconti1529/ · write-up
Catherine de Médicis (Chenonceau) → Villeroy (in Guyenne with Anjou), BnF fr. 15564 ff. 30–34, cipher block on f. 33r (DECODE R4157, catalogued as "unknown sender (Villeroy) to unknown recipient, 5 Feb 1587"; catalogue item 173, class B)	7–8 Feb 1581	22 Sept 2026	complete	(all 289 signs accounted for: 262 read as sense, 27 nulls and padding). DECODE's sender, recipient and date are wrong: the docket reads Double de la lettre de la Royne mere du Roy à Mr de Villeroy … febvrier 1581, and the letter is printed in Lettres de Catherine de Médicis t. VII pp. 349–353, the cipher silently replaced by the contemporary marginal decipherment. The eight lines read with Satoshi Tomokiyo's reconstructed key (Cryptiana, fr. 15564 f. 33), split into finer sign classes and extended by twelve values: 18 = dict, 4/b = nous (not vous), C̄σ = mais, ß̄ = la, ō = Angleterre. The prince Dauphin to take the Low Countries' oath of fealty for Anjou, Condé to lead the forces, the embassy to England. Tomokiyo's unread last line is d'Angleterre followed by 20 padding signs. Open (graded M): ō = Angleterre, the e of Cond[é], the missing r of pouoir	targets/r4157/ · write-up
Vincenzo I Gonzaga, Duke of Mantua → Louis de Gonzague, duc de Nevers, BnF fr. 3979 ff. 92–93 (DECODE R4176, which calls the sender "duke of Nevers"; catalogue item 188, class C)	17 Sept 1590	21 Sept 2026	complete	(95.1% of 923 cipher tokens, measured). Unseparated two-digit figures read with the key in Nevers's papers, BnF fr. 3995 f. 64 "Per Cavare 1590" (Tomokiyo's no. 35), applied unchanged; beam decoder with the it-cinquecento model for the writer's 3/9, 0/8, 7/8 look-alikes. No prior decipherment. The rumour that Sixtus V was poisoned by the Spaniards, hope for Henri IV's conversion, Mantua's devotion to the French crown against the common enemy. Open: four short stretches of writer's enciphering slips (about 45 tokens), the same on the DECODE scans and the Gallica microfilm.	targets/gonzaga1590/ · write-up
Lebel → Charles Emmanuel I of Savoy, three letters from Savoy's agent in Paris, BnF fr. 3983 nos. 11, 62, 100 (ff. 26, 130, 194–195; catalogue item 18, class C)	17 Jan, 10 & 27 Mar 1593	21 Sept 2026	complete	(≈98% of enciphered tokens valued; 21 Sept 2026). Tomokiyo published the key (savoy.htm, 2017/2023) and these letters as images only, "many code groups yet undeciphered"; no reading existed. Read sign by sign in three passes with his key, ~12 letter-sign corrections and code values harvested from the letters deciphered at the time (fr. 3983 no. 80, fr. 3984 f. 182, fr. 3985 f. 23) and from context in BnF esp. 336 nos. 79–80: no. 11 ~714 of ~720 letter signs and all 58 code tokens (41 H, 11 C, 6 M); no. 62 recto 607/617 tokens (98%), all 62 codes (36 H, 13 C, 13 M), verso ~70/101 under bleed-through; no. 100 all 1,000 tokens (codes 65 H, 26 C, 19 M). Mayenne wants the crown or the regency; Spain pays 40,000 écus and assignations to 260,000, offers four millions; the Cardinal de Bourbon offers his daughter's marriage; Mayenne's council le [archevesque] de Lyon (Épinac), Villeroy, Jeannin; the Infanta plan judged impossible; Feria's jurist mocked; a truce proposed; money through Zamet; Feria and Tassis hint Spain may turn to Savoy. Open: no. 62 verso (~6 signs under recto ink, one blot, codes 223, 247, an unsegmented 8-sign run) and M-grade code values (no. 100: 22, 27, 49, 50, 104, 150, 153, 230, 274, 278, 302, 305, XC)	targets/lebel1593/ · write-up
Ambrogio Bizozola → Maximilian Sforza (in France), BnF fr. 3034 ff. 156–157 no. 69 (DECODE R4224; catalogue item 176, class C)	4 Nov 1529 (Bologna; date line and Charles V's entry)	21 Sept 2026	complete	(97% of signs) with George Lasry's 2023 key (shown by Tomokiyo, no text published) from the Gallica scans, all four pages: the Emperor's council discussing a partition of the Milanese state (Savoy, Monferrato, Mantua, Ferrara), the eve of Charles V's entry into Bologna; homophones a/u, f/s, t added, ꝏ a line-end filler; the Archbishop of Bari's view that Milan should go to Maximilian; only the date numerals and ~35 signs doubtful	targets/bizozola1520/ · write-up
[Henri, duc de Guise?] → Philippe-Emmanuel, duc de Mercœur, BnF fr. 15564 f. 78 (DECODE R4158; siblings R4155 f. 27 read by Lasry, R4167 f. 151 = catalogue 13; catalogue item 171, class B)	16 Apr 1587	22 Sept 2026	complete	(983 of 998 cipher signs) with George Lasry's 2022 fr. 15564 key, extended by about 15 values (he published the key and f. 27 only): Mercœur asked to stay in Brittany until Mayenne(?) takes the place after Bellièvre arrives; Parma's troops cannot stay; "23" to be settled par la douceur; a lady who wants pensions paid par la reddition des places. Three code groups, 12 signs and the name codes 8ı, 23, 12 N open.	targets/mercoeur1587/ · write-up
[Henri, duc de Guise?] → Philippe-Emmanuel, duc de Mercœur, BnF fr. 15564 ff. 119 and 142 (DECODE R4162, R4165; siblings f. 78 = mercoeur1587, f. 27 read by Lasry; catalogue item 172, class B)	27 May and 20 Jun 1587	22 Sept 2026	complete	(95.1% of 2,477 letter signs, measured) with George Lasry's 2022 fr. 15564 key, extended by about a dozen values (he published the key, not these texts): two thousand Italian or Albanian lances and four thousand reiters for August; a secretary to Colonel Pfyffer with 12,000 écus for nine thousand Swiss; the Protestant army of 8,000 reiters, 12,000 Swiss, two lansquenet regiments and 4,000 arquebusiers for Châtillon; by June 'nous sommes maintenant contraincts de donner ordre a nos affaires', forces to Meaux. Name codes 21, 43, 71, 73 and 121 scattered signs open.	targets/guise1587/ · write-up
Cardinal Marino Ascanio Caracciolo (Milan) → Charles V, AGS Estado leg. 1184 fol. 110 (DECODE R9966, "Non-decrypted"; catalogue 196, class C)	14 Nov 1537	21 Sept 2026	complete	(96% of plaintext letters, measured; 21 Sept 2026). The eleven cipher lines read with Wanruo Luo's Cifrario 38 (thesis 2021, the Charles V–Caracciolo cipher of 1530–38) unchanged; Tomokiyo had named the key and read the first four lines (two words corrected here). Content: how Marnol was to treat with the Swiss and the Grisons, Panizone sent and paid in Caracciolo's name, terms for the Grisons at Como, Marnol now aware of "la malignita et perfidia" of the Swiss and their ministers. Second pass: superscripts belong to the preceding sign, bare ♀ = r, barred g = f; the Grisons to get "la tracta del frumento gratis" as the Swiss had. About ten signs doubtful	targets/caracciolo1537/ · write-up
Wolff (The Hague) → Wilhelmina of Prussia, Princess of Orange (London), KHA The Hague, Wilhelmina Prinses van Pruisen inv. 192 (DECODE R2232, "non-decrypted", "homophonic"; catalogue 222, class A)	25 Jul 1801	21 Sept 2026	complete	Read with the key R2233 Wolff sent on 8 Sept 1801. 45 rows of four consecutive values; number + superscript 1–4 picks one (up to four homophones per letter). A 1798 invasion plan (Le Havre, Paris bombarded, Holland, the IJssel) sent to the Princess and the King of Prussia; offer of Swiss troops to Prince Frederick cut short by his arrest on 14 Sept 1795 and 39 months in prison; plea for British half-pay. A few groups one place off; two clear passages summarised; Wolff unidentified	targets/r2232/ · write-up
An Orangist correspondent (Basle) → Prince William V, KHA The Hague, Prins Willem V inv. 339 (DECODE R2234, "non-decrypted", "homophonic"; catalogue 221, class A)	17 Aug 1796	21 Sept 2026	complete	Read with the key sheet R2235 from the same file. Five alphabets D, G, L, N, R, one per line, cycled from each section's capital (not homophonic). Lodging and pay at Basle, Salm's hussars for Prussia and Orange, Osnabrück, Barthélemy and his secretary, Degelmann, the armies of Italy and the Rhine. A few words doubtful from transcription; clear signature unread; writer unidentified ("Tollius" in pencil)	targets/r2234/ · write-up
Pseudo-Johann Arndt, Das zweyte Silentium Dei, an alchemical manuscript book, Glücksbrunn, Beinecke Mellon MS 136, pp. 24–76 (DECODE R2873, "Non-decrypted", "graphic signs"; catalogue 168 'Johann Arndt to unknown recipient', C)	18–25 Sept 1798	22 Sept 2026	complete	Read with the key drawn in the book (p. 83). Not a letter: a Rosicrucian-Paracelsian book on a burning-glass "magische Machine". Pigpen-type "geheime Schrift" with ligature signs; all 31 passages (1,545 signs, 345 words) read, 2 signs unread (a vessel sign outside the key, a gutter word), 4 words doubtful. p. 24 was already read online by Lasry/Kopal (Schmeh, Cipherbrain, March 2022); five short passages read here for the first time, earlier readings corrected	targets/mellon136/ · write-up
Arend Willem van Reede (Berlin) → Stadholder William V, KHA The Hague, Prins Willem V inv. 205 (DECODE R1057, "partially decrypted"; catalogue 217, class B)	4 Feb 1792	21 Sept 2026	complete	Deciphered with the Grand Chiffre key R1024. A cover note: M. de Larrey is told there is lemon juice (invisible ink) on the four blank half-pages of the enclosure and asked to pass the letter to the Prince; the last 21 groups are declared nulls. One group (186 4115) doubtful; the enclosure is not on the record	targets/vanreede1792/ · write-up
Cornelis Haga (Constantinople) → the States General, two ciphered despatches, Nationaal Archief 1.01.02 (Staten-Generaal) inv. nr. 6894 (DECODE R2113 and R2115; key R2118) (catalogue 229, class A)	1 Feb – 4 May 1620	21 Sept 2026	complete	DECODE's “zodiacal and alchemistic signs” are a plain letter substitution whose signs are letter shapes, with a 116-entry numeric nomenclator; the key for the series is in the same inventory number (R2118) and fits at first try. The two records hold four letters, not two, and their pages are bound out of order (R2113: the clear close of 1 Feb, then a letter of 31 Mar; R2115: 24 Apr and 4 May). The cipher carries Haga's advice on the Bohemian revolt — the Hungarians to elect regem Bohemiae and crown a new king, Transylvania to be resigned to Bethlen's brother, orders to the Bassa of Buda — and the Porte asking whether the truce with Spain will be renewed, since it would rather see Spain attacked from both sides. The key's m cell is off the photographed sheet and was recovered from the letters. Code (50) does not fit its key line; codes 37–45 rest on inferred numbering; a few short cipher words open	targets/haga1620/ · write-up
Charles I and Nicholas → Boswell, TNA SP 84/157 ff. 217, 219	1643	15 Sept 2026	complete	Alphabet solved by R. Pitt (Sept 2026): a 24-letter row repeated four times over 20–115, plus supplementary homophones and nulls; verified here against a 20,000-row permutation control (z = 9.6). Added here: the four graphic signs read as word-signs introduced inline (good, Cousin, Master, us), which closes the framed passages; the King's letter is to the Duke of Courland's envoy, sent via Boswell with the re-credentials Simpson printed in 1893; Nicholas's letter concerns the Dutch embassy of 1644. Both letters read in substance; a dozen single-occurrence word codes remain, and the transcription needs the folios	targets/boswell/ · write-up
Louvois and Louis XIV → Catinat, seven despatches, 8 July – 14 Sept 1691 (Mémoires de Catinat 1819, t. II pp. 295–342)	1691	16 Sept 2026	complete	12,362 groups of the Grand Chiffre de 1691 decoded in full with Bazeries' 1893 table from the MDZ hOCR of the volume, checked against the page images; the two letters Bazeries printed in clear agree at 98.1 % and 99.1 % of characters, and the other five (9 July, 24 and 29 Aug, 6 and 14 Sept; about 9,700 groups) are read for the first time in print. The 14 Sept letter is the King's decision to bring the army back over the Alps, fight defensively in the passes in 1692, hold Carmagnole to cover the negotiation with the Pope and then burn it, and take Coni in winter. Six groups misprinted above 587 in 1819 stay unread	targets/catinat1691/ · write-up
Henri IV → Landgrave Maurice of Hesse-Kassel, seven ciphered passages (Rommel, Correspondance inédite 1840, pp. 84, 99, 211–214, 230, 269–272, 307–314, 391)	1602–09	17 Sept 2026	complete	(17 Sept 2026): about 4,100 groups transcribed by eye from the Internet Archive page images and decoded with the key Rommel himself printed in 1846 (Allg. Zeitschrift f. Geschichte V, 402–403), with no published application to them; the texts fix the misprinted lists (70 = p, 85 = m, 83 = o, the name row shifted by one, 59 with overbar = le Turc). Control: Rommel's 1846 clear text of the 3 April 1604 letter. 22 Dec 1605: Mérargues and the Gunpowder Plot "forgés sur mesme enclume", two million livres for the Dutch again in 1606, money and troops asked of the German princes before spring; 20 May 1606: the Spanish design on a King of the Romans and the call to the princes to "prendre conseil ensemble … pour la conservation de la liberté germanique". A prince's name on p. 213, seven groups on p. 307 and one unlisted name-group remain	targets/hesse1603/ · write-up
Sormano and de Vaulx → François I, Ferrara, 23 Feb 1529 (nos. 65, 66) and the sequel no. 63, BnF fr. 3096 ff. 115–122 (catalogue item 8, class A)	1529	17 Sept 2026	complete	(17 Sept 2026): nos. 65 and 66 end to end, no. 63 as far as its leaves go (it ends mid-sentence on f. 116v; the last leaf is not at f. 117). Lasry's 2023 key verified on the glossed sibling no. 62 and applied at glyph level: 18,273 signs segmented from the Gallica scans, k-means clustered and hand-labelled, k-NN classified, then every line checked by eye on review sheets. Not in the published table and proved on the leaves: a null, the crossbarred long-s = e, a double-circle nomenclator for duca, ꝓ = d, ⊡ = ll, ≠ = cc. No. 66 (docketed duplicata) enciphers different stretches from no. 65, so the pair cribs itself. Alfonso d'Este risolutamente concluse per cosa dil mondo non voler per sé alcunamente accettare il regno et manco far l'impresa a suo nome; he will accept only the accord the King makes for him with the Pope, pleads the want of provisions, and the agents fall back on Ercole d'Este and Trivulzio	targets/sormano/ · write-up · second write-up, by Arya Sanketbhai Patel
Gramont, Mâcon and Langeac → Montmorency, Venice and Rome, 1529–1537, BnF fr. 3083 no. 8, fr. 3091 no. 23 (ff. 45–47), fr. 3071 nos. 4 and 7 (catalogue item 6, class C)	1529–37	18 Sept 2026	complete	(18 Sept 2026). Four letters, four keys (not three), all reconstructed by Lasry in 2023 (tables on Tomokiyo's GL.htm); Lasry decoded all four texts at the time, but the decipherments were not posted. The work here is a new public transcription and reading. Langeac 14 Jan 1529 (Saint-Pol and the Signoria) and Gramont 11 Oct 1529 (171 lines of solid cipher: Clement VII on the general council, feigning illness for the Admiral Chabot, soothsayers and a dream before Bologna, Orange's Florence deal) read in full; Gramont 21 July 1530 ~85 %; Mâcon 11 Apr 1537 ~90 % of the cipher (Farnese marriage, Paul III's threat to declare for the Emperor, the Turk). Code signs identified: heart = King, 8Δ = pape, 20/30/40 = Empereur/pape/roy. The Mâcon slip f. 9bis is a contemporary decipherment and agrees	targets/gramont1529/ · write-up
Robert Yard → the Earl of Manchester, Whitehall, 12 and 16 Oct 1699, Beinecke OSB MSS fc37 box 2 folders 49, 51 ("almost entirely in cipher, undeciphered")	1699	18 Sept 2026	complete	(18 Sept 2026): the key was in the same Manchester papers at Yale, the "Diplomatic cipher, contemporary copy" (box 18/47, OID 2046948): a printed one-part template, 14 columns A–Z, with handwritten numbers 3–2342 that skip every number ending in 5 or 9 and the range 516–617; those are nulls ("Blanks"), and so in practice are the unassigned slots. Transcribed (1,456 slots) and validated on the contemporary interlinear decipherment of the 5 Oct sibling; it then reads all 368 groups of both letters, apart from five slips of Yard's pen. Yard's copy differs in three values: 42/48 = e, 21 = d (as HMC Addendum 141a had it), 483 = and. Tomokiyo had identified THE=454 and decoded the first groups. Content: messengers at Dover for Mills and Lord Drummond's priest; a design "carryed on very privately at Saint Germains"; a Paris letter for Mr Gray opened at the post office and resealed. Cole's Memoirs (1733) not re-checked	targets/yard1699/ · write-up
Mary Queen of Scots → the Duke of Norfolk, "the 20th" [Feb or Mar 1570], BL Cotton MS Caligula C II f. 74r	1570	18 Sept 2026	complete	(18 Sept 2026): all 23 lines. The only letter of the Norfolk series with no contemporary decipherment; Tomokiyo marked it undeciphered (no contemporary decipherment, not unsolved), rebuilt the key from ff. 66/69/75 and overlaid a partial letter-by-letter decode. Key checked on f. 66r against its decipherment, then applied to new BL IIIF images stitched from 1,024-px tiles. Mary asks Norfolk to write through the Bishop of Ross or "your sister Scrup" and not in his own hand; Elizabeth "sueves me the veyt of Murray death"; "the brut is that Murton is gon". Open: five unkeyed nomenclator signs, two groups in line 1, a stain at the end of line 4, single words	targets/norfolk1570/ · write-up
Galeotto Fibindacci da Ricasoli → the Signoria of Florence, Urbino, early 1425 (1424 Florentine style), ASF Dieci di Balìa, Responsive 2 no. 171 (DECODE R3754, "non-decrypted"; oldest-cipher search)	1425	18 Sept 2026	complete	Key rebuilt by Gabbrielli (1863) from a few glosses and printed by Meister (1902, p. 50), but the letter never printed. DECODE's "non-decrypted" is wrong: the interlinear glosses are a decipherment; what the record lacks is the key or a reference to it. The three long runs Gabbrielli left unread decoded sign by sign with his glosses as seeds: Madonna è pure femina e sospetosa (the Count on Caterina Colonna), parenti e citadini bind themselves in secret to the Count, acomandino a lui. His null list corrected (ɤ = A; q confirmed null); four new values (ɤ, +o, 9, ô) confirmed on siblings nos. 21 and 91. Open: two word signs read from context (il Papa?, ella), five letter values on this letter only, some words of the clear hand	targets/florence1429/ · write-up
Charles VI → Count Leopold Viktorin Windischgrätz (The Hague, Cambrai), four autograph letters, Vienna 3 Feb 1720 – 11 Apr 1722, SOA Plzeň RA Windischgrätz inv. 694 and inv. 5 book 5 pp. 401–416 (DECODE R5019–R5024, "non-decrypted"; catalogue item 69, class A)	1720–22	19 Sept 2026	complete	(19 Sept 2026) with the two keys in the same family archive, DECODE R5017 (Chancellery nomenclator, 1–977) and R5018 (the Emperor's private key, numbers and signs), both transcribed in full; no cryptanalysis. All 35 cipher passages read: Alberoni's fall and the fear of a French separate peace with "the Duc d'Anjou" through Morville, Sardinia for Lorraine "sehr ideal", the Montferrat equivalent, Portugal into the Alliance (1720); Orléans and the Barrier burden; Beretti Landi's proposal of a marriage between the Emperor's children and Philip V's sons, "beede so klein" (July 1720), and again in 1722. Two nomenclator cells uncertain (446, 638); key slips 31 = zu, 268 = ein	targets/windischgraetz1720/ · write-up
Cardinal Alessandrino, Rome → the nuncio in Spain (Castagna), 6 cipher letters, AAV Segr. Stato Spagna 1, ff. 165–324 (DECODE R85–R90, "partially decrypted", catalogued as unknown sender to unknown recipient, 1 Jan 1566; catalogue item 261, class B)	1567	21 Sept 2026	complete	(21 Sept 2026). Sender, recipient and year from the archive's description of the volume (ff. 146–338 = Alessandrino's letters and ciphers to the nuncio, 1567). Lasry's 2020 key on the records, identical to the one on R93–R102, reads all six with the alessandrino1568 decoder unchanged: Philip II's journey to Flanders and a papal interview, Borromeo against the Milan senate, the unconfirmed archbishop of Cologne, and (R88/R89, two copies; R90 its close) advice to imprison Montmorency and his son until Condé and Coligny disarm, Oct–Nov 1567. Open: the group 'colospo', 8 dotted groups outside the key, Serrano 1914 not checked	targets/nuncio1566/ · write-up
Cardinal Alessandrino, Rome → the nuncio in Spain (Castagna), 11 cipher letters, AAV Segr. Stato Spagna 6/I items 1–10 and 6/II item 13 (DECODE R93–R102, R115, "partially decrypted"; catalogue items 253 and 254, class B)	1568–69	21 Sept 2026	complete	(21 Sept 2026). Lasry's 2020 reconstructed key was on the records but no plaintext: a polyphonic cipher, each digit two letters (1 u/t, 4 a/l …), 0 null, X0Y0 nomenclator. A beam search choosing each digit's letter under the Italian model reads all eleven end to end: a French bride for Philip II, jurisdiction and the King's ministers, Bolognese exiles, the Turk and the German Protestants, Borromeo against the Milan senate. Open: 15 dotted groups outside the key, single-letter choices, comparison with Serrano 1914	targets/alessandrino1568/ · write-up
Paris nunciature (bishop of Bergamo, archbishop of Nazareth; Morosini) → the Secretariat, 3 ciphertexts, AAV Segr. Stato Francia 18 (DECODE R15–R17, "partially decrypted"; catalogue item 242)	1585–87	21 Sept 2026	complete	(21 Sept 2026). R15 was read by Lasry on DECODE. R16's DECODE decryption is an md5-identical copy of R15's, so R16 (3 Mar 1586) was unread: read with Lasry's F18 key plus the nomenclator Meister printed (p. 324, Ragazzoni 1583): Guise refuses the sale of church property under the bull unless given Metz and money; the clergy's appeal to a future council. R17 read with Meister no. 40 (1587) via the morosini1588 decoder: Lorraine, the Guises, the reiters. Open: ~18 groups of R16 (names 18/48 unconfirmed), 5 of R17	targets/bergamo1585/ · write-up
Cardinal Gian Francesco Morosini, legate in France (Blois, Tours, Moulins) → Cardinal Montalto, 36 cipher enclosures, Aug 1588 – July 1589, AAV Segr. Stato Francia 22 items 1–37 (DECODE R18–R22, R27–R49, R51–R58, "partially decrypted"; catalogue item 243, class B)	1588–89	19 Sept 2026	complete	(19 Sept 2026). The key is Meister 1906 no. 40 (issued to Morosini 8 June 1587); DECODE had it and Lasry's partial reconstruction but no plaintext. The figures run on without separators and homophones are 1–3 digits, so each page was segmented by a beam search scored with the shared Italian model. 34 enclosures read end to end; R53–R54 (spring 1589) are in a second syllabic key known only in part and read in part. Content: the fight to publish Trent in France, the Estates of Blois, Savoy's seizure of Saluzzo, the murder of the Duke of Guise in the King's chamber (23 Dec 1588) and of the Cardinal the next morning, the legate's disgrace, Henri III's march on Paris (June–July 1589). Open: ~1 dotted code per page outside Meister's list, letter slips, most of R53–R54, item 22/29 not on DECODE	targets/morosini1588/ · write-up
Filippo Acciaiuoli, papal nuncio in Lisbon → the Secretariat of State, eight cipher leaves, 12 Sept 1758 – 1760, AAV Segr. Stato Portogallo 117 items 1–8 (DECODE R25, R195–R201, "partially decrypted"; catalogue item 240, class A)	1758–60	19 Sept 2026	complete	(19 Sept 2026). Lasry's reconstructed key (2020) had given a letter stream with all 107 code words open and no Italian text; the stream is read here as Italian and 25 codes fixed from the three leaves deciphered at the time (Re, Regina, Spagna, tutto, molto, ministri…). Content: the nuncio's secret account of the attempt on José I of 3 Sept 1758 (Teixeira wounded to death, the king hit in the right arm, taken to the Real Barraca), the wound, Carvalho's hold on the king, and in 1760 the Princess of Brazil's marriage. Open: ~80 rare codes, transcription slips in R200/R201	targets/acciaiuoli1757/ · write-up
Cesare Alberico Lucini, papal nuncio in Madrid → the Secretariat of State, clear letter with a ciphered last sentence, 13 Oct 1767, AAV Segr. Stato Spagna 304 (DECODE R120, "partially decrypted"; catalogue item 246, class B)	13 Oct 1767	21 Sept 2026	complete	(21 Sept 2026). DECODE carried Lasry's reconstructed key (2020, from sibling R121) but no decryption of R120's own passage. Read here with it after one transcription slip was fixed; 63 corrected from di to se (which also mends Lasry's R121 text), 03 = chi, 8061 = qual, 8099 = zion, 8700 = anno. The nuncio thinks Madrid's demand for a permanent licence to try clergy before lay judges aims at i tumulti interni della monarchia … e sollevazioni che succedono in diversi [luoghi]. Open: code 8107; the sentence continues on an unimaged leaf	targets/lucini1767/ · write-up
Cardinal Lorenzo Campeggio → Francis I (articles sworn to the cardinals Bourbon, Lorraine, Tournon), BnF fr. 3081 ff. 41–42 (DECODE R2322; catalogue 195, class C, catalogued as "French marshals to the King, 1 Jan 1520")	c. Oct 1534	21 Sept 2026	complete	(21 Sept 2026). f. 41r read by George Lasry (key dated 24 May 2022, on cryptiana); f. 41v, unpublished, read here with his key (99.9% of letters; Y, P, R and one 'unknown' sign re-valued): Francis will not join a crusade until Milan, Asti and Genoa are restored, and Campeggio swears, if elected pope, to press for their return. Not 1520 and not the marshals: Tournon cardinal 1530, Campeggio d. 1539, so the 1534 conclave. One sign and one code group open (99.9% of letters given a value); page breaks off	targets/marechaux1520/ · write-up
Urquhart's Cyphral Octastich, The Jewel (1652), last leaf	1652	16 Sept 2026	partial	Book cipher on the book itself: number k → first word of the needed initial on physical page k; verified from EEBO-TCP A95749 and a new 285-number transcription from the 1983 edition's photographs (HCPortal), without the plaintext Vals AI published in Aug 2026: 238/284 first-occurrence hits vs 0.43 shuffled and random-page controls. A royalist prayer for Charles II in ottava rima, Great Lord, mantaine that regal familie … Our Emperour, King, Monarch and Protector. Amen, so be it. Line 5, left open by Vals, decrypted 5 Oct 2026 as And [?] overthrow his usurp'd authoritie: 274/285 numbers give the required letter; one letter of line 5 open (p. 149), one on a leaf the TCP lacks (p. 158). Distich (Works 1834 p. 417): Proquiritation rule 34/64, chance, no English; open	targets/urquhart/ · write-up
Nicholas Throckmorton → Elizabeth I, Paris, 10 July 1559, with the margin cipher on the same leaf and John Wood → Cecil, 6 Sept 1568, BL Add MS 4136 ff. 32–33 (DECODE R2988, R2989; catalogue item 97, class A)	1559, 1568	21 Sept 2026	partial	(21 Sept 2026). Throckmorton's own letter of 10 July 1559, not in Forbes or CSP, read with Tomokiyo's Cipher 1 plus seven corrections: lines 1–4 secure, 5–15 ~85%; he sends a gentleman to report the French King's state and asks credit and reward for him; 12 code signs open. The two unidentified ciphers stay unread: the margin is a word-sign code (222 groups, 141 types, 113 used once), probably a copy of the 8 July despatches, needing its key or SP 70/5; Wood's 39-sign line has no unique solution (92 distinct fits) and needs the Hatfield original.	targets/wod1568/ · write-up
Bernardo de Salinas (London) and Pedro Ronquillo (Nijmegen) → Baltasar de Fuenmayor (Copenhagen), twenty letters, 21 Aug 1676 – 19 Mar 1678, AGR Brussels SEG inv. 2559 (DECODE R966–R984, R1001; catalogue items 37 and 41, class A)	1676–78	21 Sept 2026	partial	(21 Sept 2026; 88% of groups, below the read bar). All twenty are in the Balbases 1677 key, which reads 88.4% of 23,463 groups unchanged. Sixteen marginal decipherments confirmed. Four letters with no margin read in gist: Salinas 15 Sept 1676 and 2 Mar 1677, Ronquillo 26 Feb and 19 Mar 1678 (the last two DECODE 'non-decrypted'): the English parliament and the break with France, the northern allies' assembly, Brandenburg in Pomerania, Hanover's intrigue, the aftermath of Ghent. Six boundary-less DECODE transcriptions re-segmented by DP. Residue is transcription noise. Open: the four unmargined letters (R970 77%, R973 86%, R982 71%, R984 86%) read only for gist; their transcriptions not yet corrected from the images	targets/ronquillo1676/ · write-up
Sir Henry Norreys (Paris) → Cecil, BL Add MS 4136 ff. 160–161 (DECODE R9251 and the foot of R9250; catalogue item 90, class A)	4 Jun 1567 – 9 Mar 1567/8	21 Sept 2026	partial	Forbes's copy of only the ciphered words of seven letters; DECODE's "9 Mar 1567" is 9 Mar 1568 (OS). No key on DECODE; Tomokiyo's Cecil–Norris table (from Cabala, 1663) reads the siblings against CSP Foreign viii: demaund, reason, ruin, remain, Master Stewarde, the lord of Murrei, Savoia … intend to ouerthrou … and England, Rothes, papystes, reaums, stayid; name signs for the French King, the Protestants, Condé, the Admiral. Shuffle control: 8 word hits vs mean 0.22, max 3 in 1,000 shuffles. The 9 Mar letter's three items: the French King's name sign (H), a name sign perhaps the Queen of Scots (I), and group 2, which reads "a t ? l", not read ("Pope" and "Catholic" rejected against Cabala 1663). Open: 4 June 1567 groups 1, 3, 4, 7, a dozen 1568 groups; originals TNA SP 70 not seen	targets/norreys1567/ · write-up
Isabella of Castile → the Duke of Estrada, Segovia 20 Aug 1503, AGS Patronato Real leg. 53 doc. 69 (Bergenroth's S.E.T.c.I. L. 4 ff. 65–67; catalogue 287, class B)	20 Aug 1503	22 Sept 2026	partial	Partly deciphered with Galende's printed key. Bergenroth (CSP Spain i 369) could not decipher it. The Tratados con Inglaterra papers are now PTR leg. 53, and doc. 69 holds three ciphered copies of the letter: the 20 Aug original and duplicates of September and 3 Oct 1503. Estrada's endorsements show that the duplicates are the leaves Bergenroth calendared as nos. 380–381 (ff. 68–69), under the ratification despatch. The Cifra general de los Reyes Católicos (Galende Díaz 1994, already transcribed for esp318) reads every code group: Ya sabeys como el rey de Francia nos ronpió los días passados la guerra en Nápoles…. The letter covers the French attack on Roussillon, Fuenterrabía, Guipúzcoa and Galicia, Henry VII's treaty duty to help, and a request for dos mil peones (not the marriage). The 3 Oct clear copy in PTR 53/72 summarises it. About half the letter-sign runs are unread: PARES images are 988 px, and a higher-resolution reproduction from Simancas is needed	targets/estrada1503/ · write-up
Conde de Miranda → duque de Feria; duque de Sessa → Diego de Ibarra; Sessa → Philip II; Ibarra from Paris, Feb–Jul 1593, BnF fr. 3983 nos. 45, 79 and fr. 3984 nos. 47, 68 (catalogue item 19, class B)	1593	19 Sept 2026	partial	The Spanish office's own key survives, with its nomenclature, in another volume. The catalogue cited piece numbers; the BnF dépouillement gives folios and the leaves were found on the images — ff. 98, 162, 108, 145 = canvases 178, 286, 198, 268. Two ciphers, not one, and Tomokiyo had reconstructed a key for each, published as images a text dump drops (phelippes2.png, spanish3mansfeld.png); no reading of these leaves was published. His table has the syllables but no nomenclature, which left every line holed. Following the entry's own "verify first: DECODE for Sessa": DECODE's only two hits for these correspondents (records 4076, 4077) are not letters but keys, both naming BnF fr. 3995 — the Nevers cipher book. Folio 97 there is the office key, complete: alphabet, syllable table (identical to Tomokiyo's, confirming it against the original) and the full nomenclature (causa ʃes, estados vul, mucho hur, negocios norî, Rey tas, Roma vos…). Diacritics confirmed on plaintext: ring = −h− (achaque), dot = final −l (bolsa). f. 98 reads "con este achaque, que siendo confidente … a costa de la bolsa … yo supongo que hava de ser … no [se] verá lo que en esto convernà hazer, que yo me remito a su prudencia"; f. 162 gives a los sucesores, ordinarias, conviene. All four decode. nos. 45 and 79 are read across every cipher line on their leaves (the bands without readings are a blank gap and the folio/docket, not cipher); no. 47 across nine of thirteen bands, in its glossed stretches where the decode agrees with the 1593 decipherer's own interlinear words, and no. 68 in the second cipher (Feria–Mansfeld key, transcribed to keyM.tsv). los testemonios que a todos levantan; los que bivimos en Roma; no puedo dexar de hazerlo, a lo menos, de que no estén ministros conformes; aguarda cavallería. Open: a scatter of groups in each letter, the later lines, and the Feria–Mansfeld nomenclature, which no published table gives	targets/sessa1593/ · write-up
Scribe's colophon of the Dominican sequentiary Uppsala UB C 513, f. 19v (Strängnäs 1517, for the Västerås convent), and the note to Anders Larsson pasted on f. 1r (DECODE R4349, catalogued as "Unknown sender (Strängnäs) to unknown recipient", "six symbols in a note"; catalogue 235, class C)	1517; note later	22 Sept 2026	partial	(125 of 131 cipher tokens, 95.4%, measured). Two ciphers. The colophon, four lines under chant notation with each vowel written as the next letter (a→b e→f i→k o→p u→x), is read in full and matches the Uppsala catalogue Bd. 5 (1992), which printed it: Anno domini 1517 presens liber sequencionarius est scriptus in conuentu strengenensi pro conuentu insulensi. The six signs DECODE means close an Old Swedish note (herbs and cinnamon bark for Anders Larsson), transcribed here: ÿ ɑ M ÿ u ꝺ, probably the writer's name; the colophon's rule and all alphabet shifts fail. Open: the six signs (too short).	targets/r4349/ · write-up
Galeazzo Visconti (Campo a Landriano) → Anne de Montmorency, BnF fr. 3034 f. 154r–v no. 68 (DECODE R4223; catalogue item 177, class C, "unknown sender")	30 Aug 1528	22 Sept 2026	partial	(1248 of 1337 cipher signs, 93.3%, measured). Sender identified from the cipher (Tomokiyo; Lasry's 2023 key table "3045f28 and 3034f154", key only, no text published); read with Lasry's key extended on the letter (ne=S, me=L, y=N/B, θ=G, far=CHE; diagonal strokes are later word dividers) and checked on the signed siblings fr. 3045 f. 28 and fr. 3096 f. 91. A council of war at Landriano, Milan or Pavia: a scheme of "Francesco" with the Venetian officers swung the Duke of Urbino and the proveditor to Pavia, and Visconti was interrupted ("uno pastone de mala digestione"). Open: numbered codes 48 (×2), 4θ, 4r, rr, qq, peu (×2) (no key survives; the crossed sign is the figure 4), a recurring final sign pair, a few one-off signs and one blotted spot	targets/visconti1528/ · write-up
Georges de Selve, bishop of Lavaur (Venice) → Francis I, two despatches, BnF fr. 3091 f. 25r–v no. 14 and fr. 3045 ff. 42–49 no. 31 (DECODE R3697, R4232; catalogue item 178, class C; the card put both in fr. 3045)	14 Sept 1535 and 5 July 1536	22 Sept 2026	partial	(2,230 of 2,481 non-null cipher signs inside French words, 89.9%, measured 5 Oct 2026; 2,248 = 90.6% by key value). Read with Tomokiyo's key table (cryptiana francis.htm, table only, no text), extended on the letters: homophones, a null between doubled letters, slash = c and l, one name code glossed "Rang…". 1536 letter read (98.5% as sense): the Emperor's design on Marseille, the French galleys under the guns of La Garde, de Vaulx's advice, La Forest's news, no tenths from the Pope; all four contemporary glosses agree. 1535 letter read in part (78.2% as sense; most open signs under the strike): an Imperial informant sent to the French court as a merchant, and a watch set on the other road. Open: 182 signs under the decipherer's strike on a film scan (glosses give most of them), a 17-sign stretch with a name (Anthon[e] de Le[yve]?), a few short groups	targets/selve1535/ · write-up
A. Pelissier → Pierre Jeannin, Burgos, 13 Sept 1592, BnF fr. 3982 no. 22, ff. 46r–50r (catalogue item 16, class B)	1592	18 Sept 2026	partial	(90.9% of plaintext words measured with doubtful words unread, 95.3% counting them, 5 Oct 2026 pass 4; 183 gaps marked): Tomokiyo's key for Pelissier's cipher (rebuilt from nos. 45/46, Madrid 27 Oct 1592) fits no. 22, and was calibrated on 1,764 signs of nos. 45/46 aligned with their clear decipherment on f. 115, which split the look-alike sign classes the first transcription had merged (4 n/t, ⊤ a/p/s/g, δ o/c/r, 56 y/c, 8 n/q). All nine pages (~18,400 signs) re-read with the split classes: 3,734 words read inside the ciphered stretches, 256 unread, 158 doubtful. Codes: 315 = the Béarnais; 310 = probably Savoy, 441 = probably France (context). Philip II grants 500 ducats; Pelissier argues against holding the Estates now and for two armies (Seine–Loire, Paris and Rouen) at ~300,000 écus a month; the case being made in France for recognising Navarre; Mayenne's authority undermined; the 50,000 écus still pending. Open: the confessor's name, a few names and single signs	targets/pelissier1592/ · write-up
Ferdinand I (Augsburg) → Giovanni Maria (Johann) Malvezzi (Constantinople), Budapest, Hadtörténelmi Levéltár 1548/3 (DECODE R366, "partially decrypted", homophonic with graphic signs; catalogue item, class B)	23 Jan 1548	21 Sept 2026	complete	The key was on DECODE itself, filed at the same shelfmark as a separate record (R367, with a second copy of the alphabet as R386). Mostly clear Latin with about three pages of cipher, and the cipher is where the politics is: the French are trying in many deceitful ways to break the Habsburg–Ottoman peace, for no other reason than that, not daring to attack Charles V with their own forces, they want it done at the Sultan's charges; Rüstem Pasha is not to doubt Habsburg sincerity. The money passage (pp. 5–6) reads almost continuously: 2,800 ducats of arrears and 1,500 for the year from 1 March, carried by the secretary Justus de Argento, plus a credit of up to 3,000 repayable at Venice against Malvezzi's note. Nulls are whole Latin words (etiam, idcirco, Porro). Second pass (5 Oct 2026): every open stretch re-read from the full-resolution images with five variant sign forms of the hand; all five runs now read, 509 of 514 cipher words (99.0%) measured as Latin sense (p. 6 foot: the troops on the border; p. 7: the captives). Open: the banker's surname after Gulielmo (signs give Semaslia), one two-sign group, two lone v signs. The letter is probably in print in Austro-Turcica 1541–1552 (1995), not consulted.	targets/malvezzi1548/ · write-up
Count Leopold Viktorin von Windischgrätz, Brussels → his brother Count Ernst Friedrich (DECODE's direction; the content fits Ernst Friedrich better as writer) von Windischgrätz, SOA Plzeň RA Windischgrätz inv. 1433 karton 202 (DECODE R5029, "non-decrypted"; catalogue item 68, class A)	18 Nov 1721	21 Sept 2026	partial	With the brothers' key reconstructed by Jakub Mírka (2023), filed as a spreadsheet on the same DECODE record; no reading of this letter was published: two numbers per letter (A = 24/36 … Z = 1/37) and a small nomenclator. 181 groups transcribed from the images; all 122 letter groups and 26 known codes read (82%); second pass 5 Oct 2026 with sibling glosses (78 Compagnie, 99 Engländer, 103 Eugen, 167 Miosch): 155 of 181 (85.6%) as sense: eine Excusation zu machen, the Prince's plan, love for [103] lost daily ja selbsten Kayser, Succession, niemahlen anzunehmen. 23 nomenclator codes (33 groups) open. The Cambrai context suggests Ernst Friedrich as writer	targets/windischgraetz1721/ · write-up
Charles Hémard de Denonville, bishop and then cardinal of Mâcon (Rome, Orvieto) → Anne de Montmorency, eight ciphered letters, BnF Français 3053 ff. 6, 16, 21, 32, 35, 40, 77, 85 (DECODE R4233, R4234, R4235, R4238, R4239, R4240, R4247, R4248; catalogue 180, class A)	Sept 1536 – 6 Apr 1537	21 Sept 2026	partial	The “unknown sender” identified and every cipher passage of the eight records deciphered, seven in part; the eighth resolved. The catalogue had eight letters to the grand maître from an unnamed correspondent at Rome. Three are signed in clear Charles e. de Mascon / E. Card[ina]l de Mascon: the volume is the despatches of Francis I's ambassador to Paul III. The cipher is the one Tomokiyo reconstructed from fr. 3053 itself and Lasry re-tabulated in 2023 (used in targets/gramont1529/), with no plaintext published anywhere, so the work was reading. Homophonic substitution on graphic signs, doubled-letter signs, two nulls and exactly three code numbers (20 = l'Empereur, 30 = pape, 40 = roy — no other number occurs). Read: R4240 ~85% (Vergerio's message to the Germans, the Pope's volte-face on the general council), R4247 letter 1 ~80% (Andrea Doria's galleys, a suspected secret treaty with the Emperor, Milan refused to the Farnese), R4238 ~65% (the Sienese exiles, a German offering Ferdinand's secrets), R4233 ~60% (the Farnese marriages, a conclave weighed while Paul III fails, a design on Milan), passages of R4235 (le dict pape vouloit entretenir le roy de mariage … pour ce s'en prevaloir vers le dict Empereur) and of R4239 (Ottavio Farnese's match with Margaret of Austria, the Florentine exile cardinals, an honneste estat en France offered to Pier Luigi Farnese against the Imperial offer of Novara), and 66 lines of R4234. R4248 is not in Mâcon's cipher at all but a second fr. 3053 key, whose full contemporary clear decipherment stands on f. 86; read from that. Six contemporary marginal or interlinear decipherments found in the volume served as controls, three agreeing word for word; one of them corrects the published table (the big curly X is Q/U, not N). A third pass the same day read every remaining page: the rest of R4234 (Pier Luigi's troops held in Parma and Piacenza to garrison Milan if the Emperor sells it; Ambroise corruptible plus que tous les hommes), R4235 P2-P3 and passage D (against a bound clear copy), R4247's second letter (Pitigliano: strike at Naples) and the 100 unread lines of R4239 (Novara offered at the Pope's command). About three quarters of the cipher letters were then read. Pass 5 (22 Sept): the readers' one barred-cross alias proved to be four or five distinct key signs; every record re-transcribed by sign shape from the Gallica native scans and decoded with the key + a French LM, checked against every clear copy (ff. 37r-v found to be the clear copy of R4239 letter 2; Dupuy 44 companion letter to the King): ~95 % of all cipher letters read, CER 3-8 % vs contemporary clear texts; R4247's 'faded' tail is sharp on Gallica; 'Lodes' is Lodi, not Rodez. Open: ~5 % ambiguous sign runs, some names; the f. 85 key not tabulated	targets/rome1536/ · write-up
Arend Willem van Reede (Berlin) → L. P. van de Spiegel and the Secret Committee of the States-General, Nationaal Archief 3.01.26 inv. 190 (DECODE R1026, R1027) and 1.01.02 inv. 6695 (DECODE R1893)	29 Dec 1787 – 24 Mar 1793	21 Sept 2026	partial	Two despatches deciphered, the third open. R1026 and R1027 are in William V's 1782 Grand Chiffre (DECODE R1024), confirmed on the control R1028 whose plaintext DECODE gives; DECODE's catch-all J+ sign hides different digits and was resolved group by group with a French 6-gram model (105 groups). 452 of 472 groups read: the Prussian refusal to guarantee the Dutch possessions overseas in the 1788 alliance, Russian and Austrian moves in the Turkish war. Some stretches stay garbled by transcription misreads. R1893 (1793) is a marked three-digit code of several thousand values with no key: not read	targets/vanreede1787/ · write-up
János Pápai → Ferenc Rákóczi II, Rákóczi's envoy at the Porte (with Ferenc Horváth to 1708), MNL OL G15 (DECODE R731, R740, R741, R757, R765, R784, R793–R796, R805, R823; catalogue 46, 47 and 50, class B)	1 Jan 1706 – 8 May 1710	21 Sept 2026	partial	Read in part: eleven of twelve letters read with the envoy's own key (R793–R796, catalogue 47, four 1706 despatches, added: 5,012 of 5,043 groups, the groups above 322 in R795 being nulls). DECODE R580 (G15 C 43/39), headed Nemzetes Vitézlő Pápai Jánosnak adott Clavis, is a Hungarian syllabic nomenclator (10–322, two homophones per letter, syllables, names) never linked to the letters. Unchanged, it gives values for 11,053 of 11,224 groups (98.5%) in R805, R823, R765, R757, R741 and R740 (the last transcribed here from the image): Imperial passage across Ottoman ground, an audience before the internuncio partly in Latin, the French and Muscovite ambassadors, and the Kiaya's advice to stop spending at the Porte. Second pass: splits, joins, nulls above 322 and 220 = a bring the numeric letters to 16,641 of 16,698 groups (99.7%). Open: R731 (Belgrade, 6 Mar 1710), six graphic-sign passages, transcribed tentatively, not a simple substitution on that transcription (a control-validated solver finds no Hungarian); 57 digit fragments; Benda's 1963 edition not online	targets/papai1706/ · write-up
Catherine de Médicis and the Court → Bernardin Bochetel, bishop of Rennes, ambassador at the imperial court: BnF français 3181 f. 55; Cinq cents Colbert 390 pp. 139, 357; Colbert 392 p. 231 (catalogue 11, class B)	Dec 1562 – 1564	20–21 Sept 2026	partial	Four letters never deciphered, read in part (79% of 2,308 signs measured as sense, 15% probable; 5 Oct 2026). Bishop of Rennes’ cipher (Tomokiyo’s reconstruction): homophonic, ~30 word-signs, clear-looking word nulls. Each letter is in a different secretary’s hand, so each hand’s forms were fixed on letters in that hand that carry contemporary decipherments (fr. 3181 f. 57–58; Colbert 390 p. 138, pp. 189/199, 221–231), and each line decoded by a language-model lattice over the candidate values (lattice.py, fr-1530-despatches). f. 55 (Catherine, 31 July 1563; La Ferrière II 79–81 printed the cipher as « Partie chiffrée »): ~91%, the Council of Trent and the Habsburg marriage — pour le bien de la Chrestienté et avancer le concile… ce qui se promect des sessions de Decembre… vous estant dernierement à … Trante; f. 57’s margin reads par France, not par Ferrare. Colbert 390 pp. 357–358 (1564, after Ferdinand I’s death; p. 358 carries the end of the cipher): ~99% — drop the precedence quarrel on another pretext, laisser là quelque secretaire… soubz couleur d’aucuns voz affaires particuliers. Colbert 392 p. 231 (Bourdin, Paris, Dec 1562; clear end on p. 232): ~85% — encloses Catherine’s letter of 15 Dec 1562 (La Ferrière I 448–451) on the secret marriage overture, which sçavoit incontinant en Espaigne; [s]a femme … triumphe de la premiere ouverture… de ce mariage. Colbert 390 p. 139 (Deschiffrez vous mesmes… et la mettez au feu): ~65%, three code words for persons/places unread — que je vous tienne pour trop advisé et affectionné et loyal serviteur du roy mon fils… les yeux ouverts pour observer… mon cousin. Calibration drew on every decipherment found in Colbert 390–395 (full ones on pp. 120–121, 199, 313). Open: signs that occur in no decipherment (f. 55 ~9%, pp. 357–358 ~1%, Colbert 392 ~15%, p. 139 ~35%, faint and null-heavy); none of the four is printed	targets/rennes1563/ · write-up
A Jacobean codebreaker's file of Italian intercepts, TNA State Papers 106 box 10, ff. 128–195 and 202–205 (DECODE R657, R660, R664, R704, R720, R721, R722, with the key R725; worksheets R658–R702; catalogue 56-adjacent, class B)	c. 1623–24	20 Sept 2026	partial	The seven records are one case, dated, and the cipher explained; the letters read in fragments. DECODE lists seven anonymous, undated, partly decrypted ciphertexts ("1558/1625", the volume's span). They are one English codebreaker's working file: intercepts with his solutions written between the lines, his worksheets, and — catalogued three records later as a key rather than a ciphertext, which is why it fell outside the target list — the fair copy of the cipher he reconstructed (R725), transcribed here in full. System A: two-figure groups read three ways by diacritic, in the codebreaker's own words, "there are numbers of woordes expressed by this difference of — over the figures" — bare = a letter (A 10/11, B 31, C 41, D 25, E 20/90 … Z 93), barred = a word of one list (16 lettere, 18 relatione, 37 Re d'Inghilterra, 52 liga), ticked = a word of a second (22 Il Principe, 23 Marques Inichiosa, 24 Imbassiatore di Venetia, 32 Roma, 34 Palatinato, 90 Papa, 91 Nuncio), plus written letters standing for whole words (a = per, g = il, q = et, z = che). Dating: that nomenclature — the Marqués de la Inojosa in England 1623–24, Cardinal Francesco Barberini from Oct 1623, Palatinato, Il Principe — fixes the file to 1623–24, against the catalogue's sixty-seven-year span. Correction to DECODE: the file holds three ciphers, not one (R657 two-figure; R660/R664/R720 three-figure; R704/R721/R722 letter-and-figure), and R720 is filed with R704 but is written in the three-figure code of R660/R664. About 45 values of the three-figure code rebuilt by aligning the codebreaker's interlinear syllables on R664 f. 1, checked by the word senza running over the line break (252 se + 359 n + 560 za); della loro sicurezza, et hoggi terzo giorno quelle de nuovo senza, and in the clear potrete continuare nel modo cominciato, osservando però il [mutamento]. Open: neither long letter read through; systems B and C not carried to a continuous reading; R657 not worked against the R725 key; no sender, recipient or codebreaker named Catalogue 62 (21 Sept 2026): seven more records of the box (R658, R661, R663, R667, R684, R701, R702) are the same codebreaker's worksheets, not intercepts; R658 is his own count-and-value grid of the three-figure code, which R701 calls "the Venetian Cifer", and it rebuilds that code as a systematic syllabary (consonant × a-e-i-o-u; 117 values, 102 in his hand), checked on his glossed strings on R702 (46/46 groups: i colonelli, sendosi, che possiamo prometter) and correcting four earlier values (260 ta, 262 ti, 263 to, 459 r); a second three-figure code and a short letter cipher on R701 stay unread.	targets/sp106box10/ · write-up
Cardinal Francesco Barberini (Rome) → Giovanni Battista Ceva, nuncio extraordinary in Paris, two letters, AAV Segretario di Stato, Francia, doss. 346 nos. 2 and 11 (DECODE R75, R84; catalogue 248, class A)	25 Sept, 18 Dec 1632	20 Sept 2026	partial	Two undeciphered letters of the dossier deciphered. The entry listed five ciphertexts as a transcription job; in fact George Lasry reconstructed the key in 2020 and read four of the eleven letters, covering three of the five, while no reading of R75 and R84 was published. System: two-figure homophones and syllables, every 2x a null, 6 a word separator occurring inside no code, and three-digit nomenclator elements (a prefix digit plus an ordinary code, not the four digits the DECODE note claims). Beam search over the unseparated stream under it-cinquecento interpolated with a model built from the dossier's own clear text; leave-one-out against Lasry's four letters, 95.8% of tokens, nomenclator recall 0.70. Controlled against the 1632 interlinear on R75: 1,105 letters, 77.5% agreement. Content: 25 Sept, the Swedish victories make l'accomodamento easier and l'unione harder, and the reach of that power sino al Mediterraneo is a danger to France; 18 Dec, a month after Lützen, the fear of conseguenze che tengano disunita per un lungo pezzo la Francia … impotente a soccorrer li suoi alleati. Six nomenclator elements read (474 mente, 495 quanto, 498 tanto, 830 Francia, 854 guerra, 149 piazza), the first solved in this dossier. Open: 117 nomenclator elements, which hold every name; transcriptions not re-checked against the images	targets/ceva1632/ · write-up
Marquis del Vasto → Charles V, Rome, early Dec 1527, BnF fr. 3022 no. 6 ff. 16v–17v (catalogue item 7, class C)	early Dec 1527	18 Sept 2026	complete	(18 Sept 2026; measured 22 Sept). No. 6, catalogued as an anonymous report, is del Vasto's own letter of early December 1527, in his Cipher 2. It is read, 98.0% of 1,680 cipher tokens measured, with Lasry's Cipher 2 key and code groups identified here from context (doz = Lautreque, deh = la Marca, dnb = dinero, bet = dello …). Lasry's syllable signs were corrected against the manuscript (boz = esto, bar = tan, bopt = ellos). It reads as his plan to secure Siena and take Perugia and Urbino. Ten code groups (11 tokens) and nine short stretches remain open. No. 10 (Cipher 1, read by Lasry 2023 and Tomokiyo 2026) and no. 20 (Ranzo's code, own row) are different ciphers in the same volume	targets/vasto1527/ · write-up
BnF Espagnol 318 nos. 5, 92, 93, 94, 95, the Catholic Monarchs' ciphered letters 1497–1504 (catalogue item 9, class C)	1497–1504	17 Sept 2026	partial	uernar*, por obra creo que … amigo, escreuir a, dificultad. Two of Lasry's signs corrected. Nos. 93 and 94 stay unread but are named: both are secretariat ciphers (homophonic sign alphabet + one-part CVC nomenclator), their code initials fall outside Escrivà's key and outside the Cifra general and inside the band of the Gran cifra of 1501–04 (lnq against Bergenroth's LUQ = capitán). The letter-run alternative was tested and rejected: 1,016 signs of f. 118r anneal to −3.5 nats/token against −1.8 for real Spanish, and 66 % of p signs sit inside pie. Blocked on BNE MSS 20.211/52, bot-checked at every endpoint and absent from Europeana and Hispana	targets/esp318/ · write-up
Nicolas Raince → Montmorency, Rome, 13 May and 20 Nov 1526, BnF fr. 2984 pp. 29–31 and p. 105 (catalogue item 5, class C)	1526	17 Sept 2026	partial	86 of about 106 lines deciphered (17 Sept 2026). Not a cryptanalytic problem: Tomokiyo published the 1526 key in 2020 with no plaintext, but reading which glyph sits under which letter in his table image by eye slips a column — the first pass here had l, m and n each one place wrong and H under r, which corrupted every reading built on it. Re-measured off the image (header letters located, ink blobs assigned to the nearest column, every one within 15 px of a centre), the table resolves a control line glyph for glyph with nothing left over: s [null] qui estoit en la court de Savoye [null] est party pour venir icy. 86 of the 106 lines then read by hand off the microfilm against numbered-token images, gaps left as gaps; 20 lines carry a machine decode only. The 13 May letter, nine days before the League of Cognac: the Castilians and the Bourguignons, deliberé prendre le chemin de Valence, a capitulation for which ilz seront cause de la destruction … de la part d'avoir faict capituler, mondict seigneur l'Admiral luy avoit dict que seroit celuy qui y tiendroit, la peste qui est au logis, des advis qui viennent aux imperiaulx par leurs, and a man taken un jour, en plaine place pres du palais, un nommé Pietro Anthonio. Its clear close, transcribed here, expects Andrea Doria at Civitavecchia within a day with six galleys. The 20 Nov letter, two months after the Colonna raid, turns on a pontificate that seroit la cause de la totale ruine de sa maison. One glyph, ten occurrences, always at a clause boundary, is in no published table. Gallica leaf numbers run two apart because they are page numbers on odd pages, not because half the volume is unfilmed	targets/raince/ · write-up
Nevers → Pisany, 8 Sept and 14 Oct 1593, BnF fr. 3985 f. 209 and fr. 3986 f. 168 (catalogue item 21; the seven "lettres avec chiffre" to Revol and Pisany)	1593	17 Sept 2026	partial	Two of seven deciphered (17 Sept 2026). Tomokiyo's Nevers cipher no. 46 (fr. 3995 f. 87, read from the full-resolution Gallica image; two-digit alphabet with homophones, barred/plain/dotted word lists, 1 as an intercalated null, checked against the office's decipherment of a Gondi letter in the same key) resolves every figure of both Pisany letters: 8 Sept reads whole ("l'on dissuadoit [le Roy] de rechercher [que le Pape] luy donnast l'absolution … qu'il n'estoyt besoin qu'il allast sans cela, parce qu'il ne feroit rien … l'on n'a volonté de contanter le Pape, que l'on n'y aille poynt"), 14 Oct in long stretches. The five letters to Revol are in the Court's symbol cipher no. 60, not no. 46: the 27 Aug copy is clear but for one line (f. 115); the cipher sits at f. 176 (2 Sept), f. 146v (7 Oct, nine lines transcribed, not read), f. 157v (9 Oct) and f. 198 (23 Oct); none read yet	targets/nevers1593/ · write-up
Unknown writer (probably Jean de Piles, abbé d'Orbais) → unnamed patron, Reims, 19 June 1589, on the Reims chapter's postulation; BnF fr. 20974 pp. 1-3, DECODE R2276 (catalogue item 194, class C; catalogued as "to the duc de Guise, 1 Jan 1556")	1589	23 Sept 2026	partial	Partial, 91.8% measured (3,600 of 3,920 signs as sense; 88.9% strict). Three pages wholly in cipher, in Tomokiyo's Nevers-Piles alphabet (Biermann 2021 identified it; no text published); the 1556 date came from another key pasted on the blank verso. Key extended from the letter (codes cardinal, leur, pour, faire, ie, vous, lettre, le/du) and from a full alignment of the deciphered sibling fr. 4715 f. 2 (~1,050 pairs). Reads the postulation of the grand archdeacon Brulart and the dean Frison, Mayenne's delays, 18,000 lansquenets against Parma, the Elbeuf ransom of 50,000 écus. Open: torn/stained edges, ~25 single signs, one code, ~80 words	targets/r2276/ · write-up
Cardinal Pellevé's secretary → Jean de Piles, abbé d'Orbais, League council in Paris, Rome, 23 Aug 1589, BnF fr. 3413 no. 62 (catalogue item 29, class B)	1589	18 Sept 2026	partial	Partial, 82% measured (22 Sept 2026). None of the three keys bound in the volume (nos. 53, 61, 69) fits. The letter is almost all in clear and carries the news of Henri III's murder. Its 91 cipher signs are in Tomokiyo's Nevers–Piles alphabet (his source, fr. 3612, is not digitised), filled out from BnF fr. 4715 f. 2, a 1586 letter in the same cipher with its decipherment written above the lines. Read: depeschera … dedans deux jours le doien; la protection; Mons. Cassin; vostre […]; Mr le [code] mon maistre a esté retiré; j'ay veu ung advis qui confirme plus particulierement que nul aultre; signature Baron. Open: three code signs and two short runs (16 signs).	targets/orbais/ · write-up
Mayenne → Henri III and the Forget–Matignon letters, BnF fr. 15572 (+ fr. 15571), catalogue item 12, class B	1585–86	22 Sept 2026	partial	Partial, in progress (28% measured): Cipher-1 key re-derived and verified on the f. 14/15 and f. 18/19 cribs and the f. 78v margin decipherment. 12 of 13 Cipher-1 leaves read in stretches (19–59% of transcribed figures each: a surrender negotiated on f. 143, the army's pay and credit crisis on ff. 154/173, a siege report naming le mareschal de Matignon on ff. 196/201, which are the same despatch enciphered twice). f. 110 is unread; Matignon's Cipher-3 leaves added by alejandrozarco (PR #25): f. 276 deciphered with Tomokiyo's key (96.2% measured), fr. 15571 f. 179 checked against Tomokiyo's interlinear decipherment (92.9%).	targets/matignon1586/
Maffeo da Treviglio → Ludovico Sforza, Buda, 22 Nov 1489 and 12 Jan 1490, ASMi Sforzesco 650 (1489 no. 8) and 642/1, 4 (DECODE R1098; Vestigia 2831, 267/268)	1489–90	19 Sept 2026	partial	Both undeciphered despatches read, the 1489 letter in part. Somogyi reconstructed the key in 2016 but published it only as numbers, with no image of any sign; her numbering follows first appearance in the letter of 2 Apr 1490, which has a contemporary clear copy, so the numbers could be matched back to the signs, and ~20 further signs were recovered from the letters themselves. The 1489 letter (four pages, of which only two lines had ever been deciphered) reads nearly throughout on pp. 1–2, in patches after p. 3 l. 19; the 1490 letter reads in gist in all ten cipher lines and is confirmed by its own clear page 2. A place-name nomenclator sign and four minor signs stay open; the share of signs read has not been measured	targets/buda1489/ · write-up
Sir Richard Forster, 13 May 1644 (Val-d'Oise 68.H.8)	1644	15 Sept 2026	complete	Solved earlier by Found already read: George Lasry, Norbert Biermann and R. Pitt (Sept 2026) had the key; Tomokiyo's page still says unsolved. Here: permutation control z = 8.8, and a blind recovery of 31 of 34 symbols from the 207 tokens once the word boundaries are used and the French model writes u for v, i for j; six matched controls read. Not a regular Stuart key but a mixed homophonic alphabet. Four slips need the manuscript.	targets/forster/ · write-up
Sir Thomas Smith → Cecil and Elizabeth I, France, BL Add MS 4136 ff. 140, 157–168 (DECODE R9236, R9248–R9254; catalogue 89 and 94, class A)	15 Oct 1562 – 29 Apr 1566	21 Sept 2026	complete	Solved earlier by Read at the time; 1563 letters printed by Forbes, Full View ii (1741); later ones only summarised in CSP Foreign vi–viii. The six records are Forbes's copies of the ciphered stretches of 18 letters. Smith's key is Add MS 4136 f. 179 (DECODE R9261; located by Tomokiyo 2024). Read here with it: 27 Jan, 13 Apr 1563 and 9 Feb 1565 in full, most others in large part, all agreeing with Forbes/CSP. Dates corrected: Avignon 4 Oct 1564 (not 27 Oct), 22 Mar 1565/6 (not 1566/7). enos = Ambassador, not a null. R9236 (f. 140) adds five letters: 15 Oct, 20 Oct and 15/17 Dec 1562 read and agree with Forbes, 4 Nov 1563 read and agrees with CSP; only the 31 Jan 1563 advertisements stay fragmentary. Open: a few dotted word signs; 12 Mar 1563 not transcribed.	targets/smith1562/ · write-up
Sir Nicholas Throckmorton (Greenwich) → James Stewart, Earl of Moray, Regent of Scotland, BL Add MS 33531 ff. 79–80 (DECODE R8348, catalogued as "... (England) to Lorde Regent of Scotlande", 26 Jul 1569; catalogue 83, class B)	20 Jul 1569	21 Sept 2026	complete	Solved earlier by Read at the time; text recovered. Interlinear decipherment on the leaf, calendared by Bain (CSP Scotland ii no. 1103), key by Tomokiyo. A water stain has faded half the gloss, so every cipher passage was read again from the cipher. The council's plan for Mary's restoration and marriage, and a call to send Lethington. DECODE date corrected to 20 July. Name signs Z, M, La unresolved; a few words M-grade..	targets/throck1569/ · write-up
Pál Bornemisza, Bishop of Transylvania → King Ferdinand I (catalogued as unknown sender → 'Paulo Borne?'), ÖStA HHStA Staatskanzlei Interiora, Chiffrenschlüssel Kt. 13 Fasc. 20 ff. 38–39 (DECODE R1212; catalogue 267, class B)	c. 1556	22 Sept 2026	complete	Solved earlier by Read at the time; key found on the next DECODE record (R1211) and every run re-read from the signs. A contemporary interlinear decipherment covers nearly every cipher run; the docket names 'Paulo Bornemisza', and pp. 3–4 are in the bishop's own voice (Kendi urged him to give up Alba for Gyalu, 'Hec sunt verba Francisci Kendi ad me in Vasarhel'). Ferenc Kendi 'ter perfidus' denounced as the sole author of the 1556 revolt. Homophonic sign alphabet, codes 34 Transsylvania, 27 Petrovics; the signs correct the gloss in a dozen places. 276/288 cipher tokens read; codes 24, 15, 10, 41·70·55·78·18, 52 not in the key (unread at the time too)..	targets/r1212/ · write-up
Sigismund Heusner von Wandersleben (Kassel) → Axel Oxenstierna, Riksarkivet (DECODE R4332 and its duplicate R3816; catalogue 208, 209)	15 May 1637	21 Sept 2026	complete	Solved earlier by Waldispühl & Kopal, HistoCrypt 2024 (key and plaintext on R4332). Found already read. R3816 is the same three scans; DECODE's 15 March should be 15 May. Key checked on a fresh page-2 transcription. Open: ~16 three-digit codes (503 probably Landgrave Wilhelm V, M).	targets/heusner1637/ · write-up
Giovanni Battista Pallotto (Vienna) → Cardinal Francesco Barberini, 28 ciphered despatches, BAV Barb.lat. 6960 (DECODE R286–R313, "Non-decrypted"; catalogue 237, class C)	4 Aug – 29 Nov 1629	20–21 Sept 2026	complete	Solved earlier by Contents in print since 1897; key found by others. Hans Kiewning, Nuntiaturberichte aus Deutschland IV/2, Nuntiatur des Pallotto 1628–1630, Bd. 2: 1629 (Berlin 1897), prints these despatches from the Roman decipherments (ASV Nunz. di Germania 119); confirmed here against the register (R286 = pp. 5–6 = Nr. 153; p. 13 = its Beilage II). The key is the key of the sibling volume Barb.lat. 6956, broken by Norbert Biermann and Thomas Bosbach in March 2018 and filed as documents on DECODE R215. George Lasry had found years ago that 6960 uses it (DECODE was never updated), and told us on 21 Sept 2026. The system is two-digit homophones plus syllable/word groups, three-digit nomenclator groups (200–999), and the digits 1 and 8 as nulls. Applied to DECODE's transcriptions of all 28 records, it reads running Italian: R286 opens ho riceuto la risposta datami in scritto, Nr. 153 verbatim. About 1,750 three-digit groups are missing from the 6956 key and stay open. This project's own attacks (crib alignment, homophonic annealing, a fixed-width corpus search over the whole edition) all failed. They did establish, correctly, that the cipher uses deterministic groups of varying length, and that mix of widths is why every fixed-width test missed. Catalogue 236 (6956) was removed as solved by others.	targets/pallotto1629/ · write-up
Segreteria di Stato (Rome) → the Nuncio in Spain, 24 ciphered despatches, ASV i. 1025 Segr. Stato Spagna 364D (DECODE R154–R177, "partially decrypted"; catalogue 249, class A)	4 Jan 1718 – 6 Jan 1720	20 Sept 2026	complete	Solved earlier by Solved and read by George Lasry, 24 Oct 2020: his reconstructed key "ASV - S364" and his segmented decipherment of the whole Spagna 364 corpus (364C/1–32 and 364D/1–24) are filed as attachments on every one of the 24 records. DECODE's "partially decrypted" and the catalogue's "noted, a transcription job" both described the contemporary marginal gloss — a few words a page, several of them wrong — not the modern solution beside it. Measured here token by token: 33,408 of 34,243 cipher groups valued, 97.6 % read. Added here: the gap was the nomenclator, where Lasry's key gives 268 of the 399 four-digit groups the corpus uses, leaving 189 groups in 715 tokens (11.0 % of nomenclator usage) without a value. The key is alphabetical in runs, so each has a bracket; with the 85,000-token corpus for context 61 groups are recovered — promozione (45 tokens), ordin-, ciò, Card.le, port-, risoluzion-, subito, vascelli, aggiustamento, giurisdizion-, tribunal-, corriere, cifra, negoziat-, governo, istruzion-, maniera, bisogn-, truppe, Vicerè — and the calendar block resolves into two parallel Gennaio–Dicembre runs (4532–4543, 4552–4563), with 4538 Luglio confirmed independently by "delli 12 Luglio passato, qual appunto fu il giorno della promozione" (Alberoni created cardinal 12 July 1717). Unknown nomenclator usage falls from 11.0 % to 3.4 %. Two corrections to the filed key: 4474 is Vicerè, not a bishop, and 4535 "Apte" is Aprile. Open: 135 nomenclator groups (221 tokens), and 722 two-digit tokens the 2016–17 transcribers could not read off the images.	targets/nunzio1718/ · write-up
Unsigned cipher slip, Brussels → the Cologne nunciature (?), ASV Arch. Nunz. Colonia 5 f. 26r (DECODE R24, "partially decrypted", dated 1 Jan 1721; catalogue 238, class B)	9 Oct 1721	21 Sept 2026	complete	Solved earlier by Read by George Lasry (key "ASV - C15", Nov 2020) and Paolo Bonavoglia (reading), both filed on the record. Checked against the image; three values added from context (18 = qu not que, 55 = gli, 67 = g?), date and place from the clear heading. A suspended curate, news wanted from Tournai, "l'affare di Colonia". Open: 45, 43, 76 and codes 9336, 9356, 9441, 9485..	targets/colonia1721/ · write-up
Segreteria di Stato (Rome) → the Nunciature in Spain (Ab. Guiccioli), 11 ciphered despatches, ASV i. 1025 Segr. Stato Spagna 423 (DECODE R26, R179–R188, "partially decrypted"; catalogue 251, class B)	14 July – 22 Dec 1736	21 Sept 2026	complete	Solved earlier by Solved and read by George Lasry, 24 Oct 2020 (key "ASV - S423" and decipherment on every record), but his key values none of the 166 four-digit nomenclator groups (708 tokens). 46 groups recovered here from context (N.S., V.S., S. Sede, Corte, Spagna, Napoli, Brevi, -mente, -zione…); text 90.6% → 95.3% read. 120 rare groups and 136 illegible two-digit tokens open. The 1736 rupture over Spanish recruiting in Rome (Acquaviva, Montemar, Molina).	targets/nunzio1736/ · write-up
Alonso Sánchez (Venice) → Charles V, and Lope Hurtado de Mendoza, 1522, RAH Salazar 9/23–9/26 (DECODE R9593–R9657, 29 records "non-decrypted"; catalogue Tomokiyo list, class B)	1522	20 Sept 2026	partial	Solved earlier by Reconstructed and published by Satoshi Tomokiyo, Correspondence in Cipher of Imperial Ambassadors Alonso Sanchez and Juan Manuel (1522), 6 Sept 2025 — alphabet, ~177 code groups and the opening lines of R9593–R9632, from the contemporary decipherment on R9605. Added here: the catalogue entry corrected (the run spans 9/23–9/26 not 9/24; every record is dated 1522, so the July 1523 treaty is not in it; nine records are Lope Hurtado de Mendoza; almost all carry a contemporary decipherment, so they were unedited rather than unread). The nomenclator is alphabetically ordered (C1×V×C3, sorted by C1,C3,V), which fills the numeral run: cab = xv, cef = xxvi and cif = xxvii predicted and then confirmed in text, resolving Tomokiyo's uncertain "[xxviii?]" as cof. Two letters outside his coverage read in part: R9653 (20 Nov 1522) — the siege of Rhodes in clear, the foot cipher "cartas de Francia de xxvi y xxvii del passado"; and R9635 (16 Sept 1522) — a whole page of cipher with a patchy interlinear decipherment, "los de esta republica", "del señor infante", "ducados allende de los otros xviii mil", "mas de xv mil ducados", glossed fuessen restituydas sus bienes. Lope Hurtado's groups take finals (tep, tap, -z) absent from Sánchez's key — a third cipher, unreconstructed. Open: 27 records, the spelled-out letter runs on R9635 f. 20r, the head paragraph of R9653 (ink bleed), Lope Hurtado's key.	targets/sanchez1522/ · write-up
Alonso Sánchez (Venice) → Charles V, May–Aug 1523, RAH Salazar 9/28 = A-28 (DECODE R9670, R9671, R9672, R9674, R9677, R9689, R9690, R9767, R9769, R9770; catalogue 147, Tomokiyo list, class B)	1523	21 Sept 2026	complete	Solved earlier by Already calendared: Bergenroth, CSP Spain vol. 2 (1866), nos. 549, 552, 554, 558, from the contemporary decipherings filed elsewhere in the volume. Seven records are those despatches or their duplicates, matched by folio. The recipient is Charles V, not unknown. DECODE's "Inline Cleartext" flag marks clear passages, not a decipherment. The 1522 key is unchanged: the code groups on five records read with targets/sanchez1522/key_codes.tsv. The three letters of 16 Aug 1523 (R9767 Sánchez alone; R9769/R9770 joint with Caracciolo) are not in the Calendar and are mostly clear. R9767 read in part: clear text in full, and every code group in the cipher resolved except pem and yol. The cipher concerns a French-leaning party in Venice and the league's money clause. The clear postscript says the Infante was not bound to the defensive league because Venice would not bind itself against the Turk. Open: the spelled runs of R9767, the cipher head of R9770.	targets/sanchez1523/ · write-up
Sir Ralph Boswell → his cousin William Boswell, TNA SP 106/5 ff. 20–21 (DECODE R413, catalogued as King Charles I to Boswell; catalogue 63)	15 Dec 1627	21 Sept 2026	complete	Solved earlier by Found already read: Mark Woodard's monoalphabetic decipherment is on DECODE (9 Nov 2021): Buckingham back from Plymouth, confident of the King, before the 1628 Parliament. Here: his key re-checked against his transcription, p = s and w = "part" read, sender from the endorsement ("Sr Ra: Boswele. 24 Jan. 1627"), date 15 Dec 1627 from the weekday. Open: the pair w x.	targets/boswell1628/ · write-up
Henrietta Maria's household (St Germain, Paris) → Charles I, six letters, TNA SP 106/10 ff. 213–247 (DECODE R785, R787, R790, R923, R924, R929; catalogue 59, with siblings 54, 55, 60)	1645–46	19 Sept 2026	complete	Solved earlier by Found already read: George Lasry's reconstructed keys (four periods) and plaintexts for all of ff. 207–247 are on the DECODE records (23 Oct 2020); DECODE's "partial" means only the nomenclator numbers are open. His key applied here to the raw transcription of f. 247 reads it. Open: bracketed name codes, the writers "120" and "316 440".	targets/hm1645/ · write-up
ADFGVX messages, Eastern Front	1918	15 Sept 2026	complete	Solved earlier by Keys published by Lasry, Niebel, Kopal and Wacker; the 22 "unsolved" residue is garbled in transmission. The 2017 thread consolidated: 9 solved, 3 partial, 10 open; Lasry's sixteenth key rebuilt; Norbert's method reimplemented and re-derives 7 pages blind; the 10 open ones resist 15 keys and a key-free attack that fails its own control.	targets/adfgvx/ · write-up
Milroy cipher telegrams (Cox 3 Dec 1862, Bascom 4 Dec 1861)	1861–62	14 Sept 2026	complete	Solved by Richard Bean with Claude Opus 5 (2026): Stager cipher No. 7, six fixed columns, the indicator word giving the number of lines. Route spot-checked here from Milroy's own copy.	targets/milroy/
Feynman ciphers nos. 2 and 3	1953	14 Sept 2026	complete	Solved by David Vierra (May 2023): two substitution alphabets alternating word by word, even-length words reversed, boundaries dropped. Re-encryption verified here, 228 of 231 positions for no. 3.	targets/feynman/
Diego López de Ayala (Brussels) → Cardinal Cisneros, AGS Estado leg. 496 fol. 22 (DECODE R9954, "Non-decrypted"; catalogue 201, class C)	30 Aug 1516	21 Sept 2026	complete	Solved earlier by Printed in 1875 from the Simancas decipherment (the endorsement: "Copiado lo descifrado", Simancas 14 Aug 1849): Gayangos & de la Fuente, Cartas de los secretarios del cardenal D. Fr. Francisco Jiménez de Cisneros (Madrid 1875), cartas XLIX and LIV, pp. 213–241, the letter split across two numbers and misdated 1517 in LIV. Found only after an independent reading here with Tomokiyo's key from R10024 (corrected: Φ=m, hook=j/x, two s variants, H· and 6· = Xebres, a code sign = emperador). The print settles the seven words left open (sábese, emperador, Benavente, engañado, en quanta perdición, Mos de Xebres oyó) and corrects about half a dozen more (lanuça, tocar, oyda, con fe, con juramento).	targets/ayala1516/ · write-up
Thomas Urquhart's encrypted poems	17th c.	16 Sept 2026	partial	Provenance objection to the Aug 2026 claim checked independently	targets/urquhart/

Read from existing decipherment

A complete decipherment already existed; the work was transcription and checking.

Target	Date	Landed	Extent	Result	Where
Hieronimo Ranzo's code: the letter to "Garbino", Madrid, BnF fr. 3022 no. 20 ff. 44r–46v, and Ranzo's letters, fr. 2988 ff. 2, 9–10 (DECODE R1894, R4450), fr. 3019 f. 73 (catalogue items 7, 170, 189, class C)	11 Apr 1528	18 Sept 2026	partial	Initial-letter code: each group is a word's initial with a number; the numbering was tested and is not alphabetical. All 1,315 groups of no. 20 and ~2,600 of Ranzo's letters transcribed here; a word annealer validated on a held-out control recovered only function words. Read in part by others: Paolo Rosson (with Claude, 6 Oct 2026, on this transcription) values 93.8% of no. 20 and 97.2% of R1894, 79.8% and 93.1% in connected stretches; George Lasry (with Claude Opus 5.5 and GPT-6 Astra, 6 Oct 2026, on Norbert Biermann's transcription of five letters) assigns values to 95.9% of non-null occurrences (coverage, not measured accuracy), contributed 14 Oct 2026. Split from vasto1527 on 14 Oct 2026	targets/ranzo1528/ · write-up
Nicodemo Tranchedini to Francesco Sforza, 28 June 1452 (Italien 2242 — Tranchedini, 1452)	28 June 1452	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 3/3 pages transcribed in part; Partly deciphered. 852 elements deciphered; transcription incomplete. Spot-checked here 13 Oct 2026: partly confirmed.	targets/tranchedini1452/ · write-up
Jean, cardinal Du Bellay to The Constable, 1555 (Français 3142 — Du Bellay, ff60–61)	1555	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 3/3 pages transcribed in part; 3/3 pages deciphered in part. 193 elements transcribed, 109 identified; remaining ciphertext not fully transcribed. Spot-checked here 13 Oct 2026: partly confirmed.	targets/dubellay1555/ · write-up
Renard to King of Spain, 14 December 1556 (Français 20991 — Renard, ff217v–218r)	14 December 1556	13 Oct 2026	partial	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: 2/2 pages transcribed in part; 2/2 cipher pages deciphered in part. 456 elements, 346 identified (75.9%). Spot-checked here 13 Oct 2026: partly confirmed.	targets/renard1556/ · write-up
King of Portugal (Eu El Rey) to The Conde, Viceroy of India, 20April1667 (Portugais 33 — Letter of20April1667)	20April1667	13 Oct 2026	complete	Deciphered by George Lasry (with ChatGPT Astra 6 and Anthropic Opus 5.5), BnF cipher research, 10 Oct 2026: Transcribed; Deciphered. 257 elements with readings; 6 readings marked ?. Spot-checked here 13 Oct 2026: confirmed.	targets/portugal1667/ · write-up
Wallachian court → a "Reverendissimus Pater": postscript against Antide Dunod SJ, PAN Biblioteka Kórnicka BK 1560, "List zaszyfrowany" (Wielkopolska Digital Library edition 343124; catalogue 348, class C, as "unknown → the Szembek family, early 18th c.")	after Oct 1688	28 Sept 2026	complete	Read at the time: every cipher run but one carries its interlinear decipherment. Key rebuilt from the glosses (two-digit numbers 12–51, 22 letters, one number each, 15 = ae); each of the 70 runs deciphered with the key alone reproduces its gloss apart from the writer's slips (436 cipher numbers). Not a Szembek letter: it follows the death of Prince Şerban Cantacuzino (29 Oct 1688) and calls the imperial envoy "Burgundus Antidius Dunod" the impostor who misled him; the living prince is probably Brâncoveanu. Open: code group 202 (no gloss).	targets/szembek1689/ · write-up
Hanover-Celle ministers → Jobst Hermann von Ilten: Hattorf (Hanover, 7 Nov 1697 and 31 Jan 1706), Heusch (Berlin, 10 Oct 1705), Elector Georg Ludwig (25 May 1706), GWLB Hannover Ms XXIII 1245:4 ff. 212–213 and 1245:7 ff. 201–204, 288–289, 329–330 (catalogue 346, class C)	1697–1706	24 Sept 2026	complete	Read at the time: every code run carries its interlinear decipherment. The catalogue's "not deciphered" misread Kalliope's Dechiffrierter Brief, and the volumes are online in the GWLB digital library. Four letters transcribed with their glosses (492 code groups, 7 unglossed), each gloss checked against the recurrence of its group. Two codes: A (1697, null 62) and B (1705–06, used for French and German, null 91). Kolbe and Barfus at Berlin (1697); the Prussian marriage of Sophia Dorothea (1705–06). Sender of the 1706 German letter corrected to Georg Ludwig (Kalliope: Georg Wilhelm).	targets/ilten1697/ · write-up
The Cardinal of Como (papal Secretariat) → Anselmo Dandini, nuncio in France, 2 cipher enclosures, ASV Segr. Stato Francia 283C, f. 52 and ff. 89–90 (DECODE R72–R73, "partially decrypted"; catalogue item 244, class A)	1580	21 Sept 2026	complete	(21 Sept 2026). R73 already decrypted by George Lasry (DECODE, 2020). R72 uses Lasry's frame and nomenclator with a different letter assignment; letter key rebuilt here by annealing and confirmed against the contemporary decipherment on the leaf. Henri III's visits to a nun, which Rome wanted stopped. Open: 4+5 nomenclator groups, 32 illegible groups in R73.	targets/dandini1580/ · write-up
Antonio Sauli and Cardinal Riario (Portugal) → the papal Secretariat, nuncio and legate, 5 cipher letters, ASV Segr. Stato Portogallo 8, ff. 92–101, 195–197, 236–238, 281 (DECODE R190–R194, "partially decrypted"; catalogue item 260, class A)	1579–81	21 Sept 2026	complete	(21 Sept 2026). Contemporary interlinear decipherments on nearly every passage; R192–R194 transcribed and decrypted with Lasry's reconstructed key (R190–R191 already decrypted by him), agreeing with the glosses; key extended (16 Inghilterra, 30 S.S.tà, 36 Francia, 40 cattolici). Piedmont plot, papal troops for Ireland, English enterprise after Flanders. Open: 20/28, 72/82, four-digit prefixes, one unglossed tail	targets/sauli1579/ · write-up
Henri IV → André Hurault de Maisse, ambassador at Venice, 3 Nov 1592 – 27 Apr 1593, BnF fr. 16093 ff. 370, 373, 406, 410 (catalogue item 17; Tomokiyo "undeciphered")	1592–93	19 Sept 2026	complete	Two ciphers, not one, and Tomokiyo had published a key for each — as images. His page carries henryiv_Maisse1.png (ff. 370, 373) and henryiv_Maisse2.png (ff. 406, 410); a text dump of it shows no table, which is why the leaves looked keyless. He marked the four undeciphered (no contemporary decipherment, not unsolved); no reading of them was published. Cipher 1 is a simple substitution on 22 letters (a=7, e=11, f=1 … z=22; ff=66, mm=67, nn=68, ss=72); applied to f. 370 it decodes the opening letter-for-letter, 81 of 81 characters, against a contemporary clear copy found independently of the key — through the old Harlay pièce numbers (224, 226, 247, 251) that Berger de Xivrey cites — in Brienne 13 = BnF NAF 6984 fol. 102v: "le cardinal de Gondy aura passé pres de vous, il y peult ja avoir quelque jours, lequel vous aura fait…". Adds an omega-shaped sign for d not in Tomokiyo's table. All four letters read (Brienne foll. 102v, 104v, 171r, 193v); f. 406 is the duplicate of f. 404, so the four leaves are three despatches; class A, not B. Open: only 81 characters of f. 370 transcribed group by group, f. 373 none, and Cipher 2 keyed but not applied	targets/maisse1592/ · write-up
Henriette de Clèves; Renato Birago → the duke of Nevers, Lyon, BnF fr. 3315 ff. 15–16, 21 (catalogue 283, class C)	16 and 28 Sept [1574]	23 Sept 2026	partial	Birago's 28 September Italian letter is complete in a clear fair copy beneath its cipher flap: Nevers's delayed visit, the ill-timed business of the bearer, Piedmont, and Their Majesties' willingness to grant him a governorship. The duchess's two cipher pages have a minute interlinear French decipherment; its opening and the clear ending on f. 16r are read, but most of the middle is not continuously transcribed. The related Nevers no. 2 alphabet of 1584 is not the exact key.	targets/nevers1574/ · write-up
Benedict XIII → Francesc Climent, and his circle to Climent, Avignon/Perpignan/Peñíscola → Barcelona, Arxiu Capitular de Barcelona, Climent's papers (Puig's "Documentos inéditos del obispo Çapera", nos. 1063, 409, 1068, 463, 461 …; Meister 1906's capsa "Documentos referentes a la familia Luna I"; catalogue 334, the top pre-1425 lead)	1399–1416	23 Sept 2026	complete	Solved earlier by Deciphered by Sebastián Puig y Puig, Pedro de Luna, último papa de Aviñón (Barcelona 1920): 14 cipher letters printed with the deciphered passages in italics (Benedict XIII's own 1399–1403, ~2,200 words; Exea 1408, Comuel 1415–16), key rebuilt by Puig from lines Climent had deciphered on docs. 1063, 461, 463, never printed. Meister (1906, from Ehrle) never saw them; the 2026 archive scan could not reach Puig (HathiTrust), found here on Google Books. Added: the 14 items listed; Puig's p. 475 facsimile of the last page of the 22 June 1400 letter found, and the key Puig never printed rebuilt from it against his text (key.md): simple substitution, one sign per letter (20 values), name sign V̇ = King of Aragon, null separator groups; 755 of 826 plaintext letters on the plate confirmed sign by sign (91.4%; 96.3% of the legible runs), none against the key; Puig corrected three times (te sicudum → tenendum, abique congerio → absque congedio, necessaria). Open: plate line 4 (overwritten) and 16 blurred words, one 1398 cipher line in an unknown key, Climent's two keys in doc. 595, letters cited but not printed; archivist's note that Puig did not return many Schism papers.	targets/benedict1399/ · write-up
Cardinal Matthäus Schiner → Henry VIII, Trent → London, BL Cotton MS Vitellius B XIX f. 131 (old foliation; f. 129 in the BL's current catalogue), L&P II 2044 (catalogue 290)	14 June 1516	23 Sept 2026	complete	Solved earlier by Deciphered by Albert Büchi and Traugott Schiess, Korrespondenzen und Akten zur Geschichte des Kardinals Matth. Schiner II (Basel 1925), no. 541: the autograph printed in full, cipher "dort nicht aufgelöst" and given in italics, keys rebuilt by the editors and never printed. L&P (1864) had called it "not deciphered"; found here on e-Helvetica after HathiTrust refused. Added: the 14 cipher stretches (88 words; 84 on surviving cipher, 4 Büchi's conjectures in the burnt margin) located on the page images and translated. They give Maximilian's proposal to Henry of the duchy of Milan and the crown of the Empire, the Duke of Bari's knowledge of it, and Galeazzo Visconti's dealings over Milan. L&P shown to print the words of 6 of the 14; the 1516 London abstract on the next leaf (L&P 2045, Büchi 542) confirms the ciphered content; foliation reconciled. Open: a sign-level check and the key, which need the leaf (not digitised).	targets/schiner1516/ · write-up
Erving → Madison, Madrid, 24 Mar 1807, private No. 21 "in the Cypher of the Legation" (Pinckney's code; Founders Madison/02-13-02-0559)	1807	18 Sept 2026	complete	Solved earlier by Every hole in the printed text closed (18 Sept 2026). The editors print Madison's interlinear decode and say the key was not found; Madison had it, and the holes are his slips. Collated on both received copies (NARA M31 reel 12 frames 0283–0286; LoC mjm014714 duplicate) and on Pinckney's decoded despatches (reel 7 frames 0346–0352, reel 8 0083–0084): 69.1651.1183.1497.1298 = in the scale (JM "to the se al e"), 724.529.934 = Hanover ×5 (JM "Haweder", "Ha——er"), 1578 = that ("see that he was more so"), 1619.1508.1497.1594 = Gibraltar (the [sic] is JM's). The only loss is Erving's own dropped -eak after 1212 sp, missing in both copies. A small completion of a text already in print.	targets/erving1807/ · write-up
Phelipeaux d'Herbault → Philippe de Béthune (Rome), 13 Feb 1626 (BnF fr. 3669 no. 25)	1626	18 Sept 2026	complete	Solved earlier by Resolved and read (18 Sept 2026): the catalogue's "one letter without decipherment among ~30 of 1625–26" is not an undeciphered text. No. 26 is not a second letter but the same letter of 13 February, carrying the same cipher with the plaintext written between the lines by a contemporary decipherer; the clear text of the two copies runs word for word, the cipher stands at the same points, and no. 25's first run opens to 92 … 17 … 43 and its third ends … 68 93 … exactly as no. 26's does. The letter reads in full: the Legate's departure for Spain and "le Pape pour l'acheminement de ces troupes in la Valteline", the fear "sy le Pape entroit dans la partialité"; the peace with the Huguenots, needed because "le faix d'une double guerre ne se pouvoit plus supporter"; the Dutch ships withdrawn and the Huguenot admiral following "avec le reste de sa flotte" without the King's permission; and the prince of Piedmont's propositions, which "tendent à engager le Roy à la guerre avec l'Espagne, (et que) la Seigneurie de Venize contribue comme faict l'Angleterre et Holande" — the Savoyard push Monzón defeated six weeks later.	targets/herbault1626/ · write-up
Philip II → Mendoza, two letters of 7 Sept 1589, BnF fr. 3641 ff. 10/14 and 12/76 (catalogue item 32, "second undeciphered letter")	1589	17 Sept 2026	complete	Solved earlier by Resolved (17 Sept 2026): the "second letter" (f. 14) is the 1589 decipherer's fair copy of the Cg.13 original on f. 10, laid out like the original with the unread nomenclator groups underlined and listed in the margin; f. 76 is the same for the other letter (f. 12), miscatalogued by Omont. Both pairs aligned group by group: a partial table of Philip II's general cipher Cg.13 (some 70 syllables in rows of five, three ending marks, c. 50 code groups), consistent with Tomokiyo's published fragment at every overlap. Cross-reading the letters reads groups the decipherers left blank (pur respeto, mes vuestro, san negocio; bur persona, dom dicho, dur porque, pen mucho from Phelippes's interlinear f. 5), shows tur to be a miscopy of tum, and corrects the copyist's "Francia" for dun (England: the fleet's return, "entre ellos y el de Bearne", "sus armaçones") and "si se hiziesse" for the King's "si yo hiziesse". Fourteen groups in the two letters still unread; the Simancas table (Devos 1950) would close them.	targets/mendoza1589/ · write-up
Prospero Santa Croce (nuncio in France) → Cardinal Innocenzo del Monte, ASV Segr. Stato, Nunziatura di Francia 3 ff. 117, 128, 230, 248, 263 (DECODE R6–R10, dated 28 Jul 1552 there; catalogue 245, class B)	26 Mar – 23 Dec 1553	21 Sept 2026	complete	Solved earlier by Read at the time; in print. Decifrati bound in the volume; all five printed deciphered in Lestocquoy, Acta Nuntiaturae Gallicae 9 (1972), nos. 67, 71, 112, 124 and 23 Dec (II); Lasry's key on DECODE since 2020 reproduces the print. R9's three cancelled cipher lines, unprinted, read here: the clear Corsica/Strozzi postscript. Code groups not valued..	targets/santacroce1552/ · write-up
Mary Queen of Scots (Bolton) → John Hamilton, Archbishop of St Andrews, BL Add MS 33531 ff. 73–74 (DECODE R8347, catalogued as "... to unknown recipient", Jan 1568; catalogue 84, class B)	18 Jan 1569	21 Sept 2026	complete	Solved earlier by Read at the time. The contemporary decipherment is bound in as f. 74; calendared by Bain (CSP Scotland ii no. 966, second of the number), key by Tomokiyo (checked on the first cipher line). DECODE's Jan 1568 is Old Style. Decipherment transcribed in full here: Mary refuses to deliver her son to Elizabeth, expects 10,000 men from France and Spain, orders houses taken before an amnesty. Short code signs (Mn, X, k …) left open, as by the decipherer..	targets/hamilton1569/ · write-up
Thomas Randolph (Edinburgh) → Thomas Radcliffe, 3rd Earl of Sussex, BL Cotton MS Caligula C II f. 277 (DECODE R4931, catalogued as 1569; catalogue 99, class A)	5 Jul 1570	21 Sept 2026	complete	Solved earlier by Read at the time. The next leaf, f. 278 (DECODE R4932), is the contemporary interlinear decipherment; Boyd CSP Scotland iii no. 339 abstracts Sussex's clear copy. A crib attack from Boyd read ~70% before f. 278 was found and agrees with it. All five cipher runs read: the Scottish Queen's party's distrust of Elizabeth, the stayed regency, Morton, the money, the soldiers' pay, Grange. DECODE date 1569 → 5 July 1570; R4930 is not a key. Three gloss words C-grade..	targets/randolph1570/ · write-up
C. A. graaf van Rechteren tot Borgbeuningen (St Petersburg) → the Griffier / States General, Nationaal Archief 1.01.02 inv. 7407 (DECODE R1039; catalogue 232, class C)	23 Sept / 4 Oct 1785	21 Sept 2026	complete	Solved earlier by Read at the time: the Griffie's decipherment ("Ont cyfferde Missive") is images 5448–5449 of the record, which DECODE transcribed as cleartext. Margin numbers key it to the 424-group cipher. Rechteren withheld his letter to Grand Duke Paul on De Swart's advice; sure all his post is opened, he asks for couriers and replies in cipher. Key not rebuilt; the pair is a known-plaintext sample of the code missing for R1040 (deswart1782)..	targets/rechteren1785/ · write-up
Charles de Poupet, sieur de La Chaulx (Vitoria) → Charles V, BNE MSS/20212/39/1-2 (DECODE R1187; catalogue 150, class B)	28 May 1522	21 Sept 2026	complete	Solved earlier by Read at the time: the contemporary decipherment is bound in as 39/2. A French letter, not Spanish, and dated 28 May, not 1 Jan. Five homophonic cipher passages marked A–E are keyed to the decipherment's sections; the first line was aligned (sign counts fit combien que je, e has three signs). Content: Adrian VI's sailing, plague at Barcelona, the Naples galleys, 4,000 men rumoured as 20,000. Key not rebuilt; C–E not transcribed..	targets/poupet1522/ · write-up
Girolamo Ghinucci, Edward Lee and Sir Francis Poyntz (Valladolid) → Cardinal Wolsey, BL Cotton MS Vespasian C IV ff. 174–177 (DECODE R8565, catalogued as "Fran. Poynts (Hispania) to unknown recipient"; catalogue 111, class B)	17 July 1527	21 Sept 2026	complete	Solved earlier by Read at the time: deciphered by Brian Tuke in 1527 and printed in Letters and Papers Henry VIII IV no. 3271. R8565 is the separated second part (old ff. 166–169) of the three ambassadors' joint despatch; placed by the folio offset from R8566 (= L&P 3270 at f. 170) and matched on DECODE's public thumbnails (clear English, about five short Lee–Worcester cipher bands deciphered between the lines, three signatures). The cipher hides Charles V's professions to Henry, Milan with the daughter of Portugal for the duke of Richmond, and De Tarbes's secret fourth instruction. Full images not viewed (DECODE login): confirm image 1 opens with the 15 July audience, not L&P 3272..	targets/poyntz1527/ · write-up
Sir Edward Stafford (Paris) → Sir Francis Walsingham, BL Harley MS 1582 ff. 74–75 (DECODE R8503, catalogued as "Sr. Edw. Stafford to unknown recipient", London; catalogue 118, class B)	19 Sept 1586	21 Sept 2026	complete	Solved earlier by Read at the time; re-read sign by sign, key values rebuilt. A holograph despatch in clear English, addressed to Walsingham. Its four cipher runs (28 signs) are glossed above: Villequier; the King of Navarre warned that men of Guise's guard were sworn to kill him and the Prince of Condé. Null 6, homophones, codes 47 Navarre, 59 Condé, 74 Guise. Not in CSP Foreign (Harley MS)..	targets/stafford1586/ · write-up
Alvise Mocenigo (Madrid) → Doge and Senate of Venice, ASVe IT ASVe 0045 010 Busta 30 f. 174 (DECODE R1874; catalogue 266, class C)	26 Oct 1628	21 Sept 2026	complete	Solved earlier by Read at the time: the chancery's decipherment is imaged in the same record. Catalogued as undated, unknown sender and recipient. Identified from the clear text and CSP Venice 21 no. 515; images 8955–8959 are the decipherment (paragraphs 6–9), 8960–8964 a second copy enciphered separately. Checked by the 6499 paragraph marks and the three 'conte di Olivares' group runs (672 2133 6210 2190 62x9 637 2127). Numerical syllabic cipher, ~2 letters per group; ciphertext-only annealing failed; the key was rebuilt later from this record's cipher and decipherment for sibling R1862 (read in part)..	targets/r1874/ · write-up
Sebastiano Foscarini (Escorial) → Doge and Senate of Venice, no. 178, ASVe IT ASVe 0045 010 Busta 30 f. 178 (DECODE R1875; catalogue 265, class C)	13 Oct 1761	21 Sept 2026	complete	Solved earlier by Read at the time: the chancery's decipherment is written on the despatch. Catalogued as '...' (Spain) to unknown recipient. Every cipher line is barred out in ink; the decipherment is interlinear on p. 1 and on pp. 3–4 for the cipher of pp. 1–2 and 5; signed and dated in clear. Content: the broken-off Anglo-French talks and a coming union with 'another court' (the Family Compact). 312 words of cipher, all read; numerical cipher, key not rebuilt..	targets/r1875/ · write-up
Nicholas Throckmorton → Cecil, Elizabeth I and the Council, BL Add MS 4136 (20 DECODE records; catalogue 88)	1560–63	20 Sept 2026	complete	Solved earlier by Archived key and prior printed plaintext verified. All 54 target image references retrieved (39 unique photographs), plus 17 images in four nearby key records. R9262 reads R9220 extract (6), “all be it i be revoked or the warr breake” (Forbes I p.355), and R9230 “orleans or burges” (Forbes II p.12): 42 tokens. Twenty-record CSP concordance; most matches are by date/addressee, not complete glyph alignment. R9255 combines other writers. Tomokiyo already identified the keys and editions; the superscript note belongs to another cipher. R2988/R2989: see targets/wod1568/..	targets/throckmorton/ · write-up
Charles IX → Philibert du Croc in Scotland, undated, Archives départementales de la Nièvre, fonds du Bourg de Bozas (DECODE R2789; catalogue 34, class C)	1565–67 or 1572	20 Sept 2026	complete	Solved earlier by Solved by George Lasry; overlays published by Satoshi Tomokiyo, 25 May 2022. The authenticated DECODE pages match Destray's 1924 facsimile and Lasry's two-page reading. Charles orders du Croc to guide the peace, government and council around James VI, prevent the young king's marriage before France is consulted, retain Scottish affection for France, and ensure “que les Anglois n'empiètent aucune chose en Escosse.” The overlay leaves several short groups open. DECODE's “Non-decrypted” status is stale, and “1 Jan 1550” is not a document date: Destray says S. l. n. d. and Charles was born in June 1550.	targets/charlesixducroc/ · write-up
Beatrice d'Aragona → Eleonora / Ippolito, ASMo Carteggio Principi Esteri, DECODE R1140–R1143 and R1154 (catalogue 160, class B)	1482–1505	20 Sept 2026	complete	Solved earlier by Earlier readings located; one record is not a cipher. MDE III (1877) nos. 10, 12 and 75 print the 1482 and May 1486 letters; IV (1878) supplement no. 45 prints the March cipher passage, with III no. 54 and Berzeviczy (1914) no. LVI supplying the other sections. R1143 has manuscript decipherment slips and repeated photographs, not five distinct written pages. R1154 is plain Italian, dated 29 May 1505, says it has no cipher, and is printed as Berzeviczy no. CCCII. Its shelfmark is b. 2/7, 11. March Virtus/Fortis/M. markers remain unexplained; no new key or exhaustive glyph validation claimed..	targets/beatrice1482/ · write-up
Frederick II → Louis Michell (London), Berlin, 28 Dec 1751, KHA The Hague, Prins Willem V inv. 198 (DECODE R1957, "non-decrypted"; catalogue 215, class C)	28 Dec 1751	19 Sept 2026	complete	Solved earlier by Printed in Politische Correspondenz Friedrichs des Grossen 8 (1882) no. 5263, from the minute. The Hague copy (Dutch interception) matched here: same clear opening, 550 groups for 438 words, repeated groups in the print's order (1064 Vienne ×4, 4310 Sardaigne ×3, 3042 dessein, 545 avec, 87 encore). Content: find out whether London knows of the Austro-Russian plan for Charles of Lorraine in Poland; report to the King alone. Cabinet code rebuilt, contributed by WAINUT (PR #29, 11–12 Oct 2026): key recovered from external plaintext, aligning six King-to-Michell letters in Nationaal Archief 1.10.29 inv. 5177 with their PC 8 concepts; the 72-value key, frozen by hash before R1957 was transcribed, decodes R1957 in the print's order (LCS 123 against a shuffle maximum of 65). Extended to 179 values, it gives 381 of R1957's 548 groups (69.5%); our 3042 dessein and 169 de anchors corrected to sur and que. The ministerial letter of the same day (R1955) is in Michell's envoy code: read in part, 95 of 360 groups (26.4%), fragments only, from the contemporary Dutch decipherments of his reports R1050/R1051. Open: the rest of R1955 (further decipherments not online; NA 1.01.50 inv. 159 not digitised); a held-out test of the extended key.	targets/michell1751/ · write-up
Queen Anne → Charles Mordaunt, Earl of Peterborough, instructions, St James's, BL Add MS 4107 ff. 184–185 (DECODE R4878, "non-decrypted"; catalogue 86, class C)	22 Feb 1711/12	21 Sept 2026	complete	Solved earlier by Printed in full in Parke, Letters and Correspondence of Bolingbroke (1798) i 310–311, from the office draft. The signed original is clear English with 61 code runs (414 groups, 242 distinct, heavy with nulls); every run aligned to the print, recurring groups constant (534 = to ×8, 678 the, 839 you, 894 him, 609 his, 1373 religion, 1219.563 Protestant, 2101 Turin). Content: not the Turin embassy but a secret mission to keep the Electoral Prince of Saxony from abjuring in Rome, and to rescue him to a Protestant country if he wished. No key on DECODE (155 British keys tested); code table not rebuilt.	targets/anne1711/ · write-up
Matthias Corvinus → Ercole I d'Este, Pozsony, 1 June 1482, ASMo Carteggio principi esteri 1622 b. 1/3 nos. 2, 2a (DECODE R1156, "partially decrypted"; catalogue 159, class B, pre-1497 candidate)	1 Jun 1482	19 Sept 2026	complete	Solved earlier by Printed in Magyar diplomácziai emlékek III (1877) no. 14 and by Fraknói, Mátyás király levelei II (1895) no. 131, from the original. Runs read here from the image: letter substitution with homophones (·ıı· = tt), name signs for the Venetians, King Ferdinand and Matthias; working key rebuilt. Four corrections to the print: valent (not "valoris (így)"), gentibus (not "servis"), the closing Speramus cito nos res nostras ita disposituros, ut meliori postea modo vos iuvare possemus (printed as nonsense), and Fraknói's "contemporary regest" is the letter's own half-ciphered last sentence. Open: faciet ꝗ venient vs. printed convenire, two name signs after regina.	targets/matthias1482/ · write-up
Letter to the King of Aragon, Barcelona, 2 May [1413–16], ACA Reserva 12 (DECODE R10170, "ca. 1420, non-decrypted, cipher type unknown"; the oldest public image on DECODE)	1413–16	18 Sept 2026	complete	Solved earlier by Read in full by Xavier de Salas, Estudis Universitaris Catalans 16 (1931), pp. 374–377, with facsimile; the DECODE record cites the article and keeps the status. Not a substitution: the slash-cut word chunks are read down the lines of each block (blocks marked by marginal dots), a columnar transposition. Rule checked on the image, block 1 reads Senyor, jo us auie quezacom escrit de so que asi s'és feit… Salas dates the content to Ferdinand I (late 1413 – Apr 1416); a royal financial officer in Barcelona with Roger de Pallars. Article PDF from ARCA in the folder.	research/oldest/ · write-up
Alchymey teuczsch, alchemical recipe book of a circle near Passau, East Bavaria, 1426, Heidelberg Cod. Pal. germ. 597 (DECODE R2876, "non-decrypted"; oldest-cipher candidate A4)	1426	18 Sept 2026	complete	Solved earlier by Read by Wilhelm Wattenbach, Anzeiger für Kunde der deutschen Vorzeit NF 16 (1869), cols. 264–268, with facsimiles and plaintexts of ff. 1r, 5v, 6v; extracts in Eis 1957/1982; the 2014 Heidelberg catalogue gives its cipher quotations resolved. The "ff. 70–71" in the candidate list were Bischoff's 1954 item numbers, not folios. Recovered here from the images: the compiler's struck 24-sign alphabet on f. 1r (a–l clear; m–ch grade C), checked on hab es and the heading niklas gank?icz (f. 11r, pencil gloss "Niklas Gankwiz"); the f. 6v invocation alphabet (19 values, repeated in abc order on f. 91v); a third set (f. 1r ll. 1–2, ff. 5v–6r, 44v, 93r) fixed for 11 letters from ist ein abc and the catalogue's f. 93r in dem caichen. Open: no continuous transcription of the recipe cipher, the f. 6r glossary or f. 44v has been printed.	targets/heidelberg597/ · write-up
Anonymous figure cipher → a League leader, 28 Oct 1586, BnF Clairambault 357 ff. 167–168 (catalogue item 14)	1586	18 Sept 2026	complete	Solved earlier by Deciphered in 1586, bound with the letter (18 Sept 2026). Tomokiyo published a partial key and left two thirds of the first page unread. f. 168 is a contemporary clear text of f. 167r down to a ✝, and from the ✝ to the end the decipherer wrote the plaintext between the lines. The key checks against the glosses (Guienne, tres confident, Fumé; adds 87 = e, 22 null, # = Roy de Navarre). Fumé, vice-admiral of Guyenne and Navarre's Catholic confidant, proposes a reconciliation between Navarre and the addressee (probably Guise). Some words uncertain.	targets/clair357/, write-up
Mihály Hentér → Ferenc Rákóczi II, Constantinople, 8 Jul 1707, MNL OL G 15 Caps. D Fasc. 80 fol. 38 (DECODE R496, "non-decrypted", misdated 1 Jan 1707; catalogue 49, class C)	8 Jul 1707	20 Sept 2026	complete	Solved earlier by Deciphered at the time, between the lines (20 Sept 2026). DECODE records no inline plaintext, but a contemporary Hungarian decipherment is written over the cipher. Monoalphabetic key confirmed on sibling letters R533/R534 (wi6xi2i4 = Felséged). The pasha of Belgrade offers troops, powder and grain against the Serbs; Hentér presses for an audience with the grand vizier; a plea for a captive's release. Some middle clauses need a palaeographer..	targets/henter1707/ · write-up
Jean du Bellay → Montmorency, London, BnF fr. 3078 nos. 3, 5, 17, 19 and fr. 3077 nos. 18, 23 (catalogue item 4)	1528–29	17 Sept 2026	complete	Every cipher passage is in print in clear. The notice hides the years; Le Grand's Histoire du divorce III (1688) prints the Court's decipherments of 15 June, 18 Sept, 17 and 27 Oct 1529 (and of 30 June 1529, fr. 3040 f. 68, DECODE R4229), and fr. 3077 no. 23 is the 28 Oct 1528 letter whose cipher Bourrilly and de Vaissière deciphered in 1905 with a key they recovered (Lasry 2022). The two passages those editions left blank, fr. 3078 no. 3 of 16 June 1529 (p. 21, DECODE R3688) and the opening of 22 Oct 1529 (copy fr. 3005 f. 165), are printed in clear in Scheurer, Correspondance du cardinal Jean du Bellay t. I (1969), no. 12 p. 41 and pp. 111–112 (seen in Google Books snippets). Done here: the key re-derived from the 22 June interlinear decipherment and Tomokiyo's table (B = o; word signs bien, fault, paix), checked on the 17 Oct letter; the 16 June block transcribed sign by sign (194 signs) and lines 0–6 deciphered from the leaf before Scheurer was found; his text agrees word for word and gave vous, faict, madame Marguerite and the last line; 191 of 194 signs have a value. Open: three signs in the last line; whether the boxed cross (madame Marguerite on p. 21, le Cardinal on p. 113) is one sign or two; the 22 Oct opening not compared with its leaf.	targets/dubellay/ · write-up
Henri IV → François Savary de Brèves, Constantinople, 1603, BnF fr. 3541 ff. 60–104 (catalogue item 23; "undeciphered as of Desenclos 2018")	1603	18 Sept 2026	complete	Solved earlier by Key and clear text both survive. The office key is BnF fr. 3462 f. 103 (Gallica btv1b9058227h, canvas 109), and its figures match Tomokiyo's reconstruction. Berger de Xivrey, Lettres missives VI (1853), printed the letters from a contemporary copy in clear (Monmerqué ms, Lettres à l'ambassadeur du Levant): 18 Feb, 31 Mar (the lists' "1 Mar") and 17 Apr in full, the other nine as summaries. Tomokiyo's partial f. 60 overlay corrected against it. Not deciphered: fr. 3541 is not digitised, so the letters have not been read against the key. A second clear copy of all twelve is BnF fr. 4029 ff. 118–128 (register of the King's letters to Brèves), also not digitised.	targets/breves1603/ · write-up
Etienne Bourdeaux (Berlin) → Maarten van der Goes, ciphered dispatch no. 16, Nationaal Archief 2.01.08 inv. 257 (DECODE R1944; catalogue 228, class C)	31 Jan 1801	20 Sept 2026	complete	Solved earlier by Read from the contemporary clear copy, misfiled under DECODE R1946. Image 3 of R1946 is the clear sheet “N°. 16, Berlyn 31 January 1801, (get.) Bourdeaux”, fol. 295, directly after R1944’s cipher on fol. 294. Rosenkrantz, sent away from Petersburg, supposes a servant copied his cipher; Lusi reports the desk holding it was forced; hence Schuer was to carry a new cipher. Open: the ~110 marked groups are not aligned and the Croiset key is not identified.	targets/R1944/ · write-up
Catokwacopa advertisements	1875	15 Sept 2026	partial	Readings audited: which the letters force, which are guesses	targets/catokwacopa/
Brâncoveanu, Cantacuzino, Turculeț(?), Chomentowski and an adviser → Hetman Adam Mikołaj Sieniawski, Wallachia, Bender, Adrianople, Archiwum Narodowe w Krakowie, Zbiór Dzieduszyckich 26/32 nr 13 and nr 5, 24/11 no. 116 (DECODE R7468–R7485; catalogue 333, C)	1707–1714	9 Oct 2026	partial	Split entry, part 2 of 2 (the Vienna-Schenck letters are under ciphertext-only). Twelve records printed by Kaczka, Danubius XXX (2012), with the Wallachian chancellery key (ten lines of R7473 checked against key, gloss and print). Six unprinted pieces carry contemporary interlinear decipherments, transcribed: R7475 (adviser, c. 1707, own alphabetical homophonic table rebuilt from the gloss), R7481–R7482 (Turculeț(?), 1709, Kaczka's key B, 99.4% glossed), R7476, R7484–R7485 (Chomentowski at the Porte, 1713–14, 3-digit syllabary; R7484–R7485 are ADzied 24/11, not "Krakow_UNK"). Token by token: R7481 99.6%, R7482 100%, R7484 96.5%, R7485 100%; R7476 (re-dated May–Aug 1714 from code XVIII = the Wallachian hospodar in Yedikule) 94.9% after seven passes (digit re-transcription at 3x, candidate words scored against the code), 67 tokens of known syllables still forming no word.	targets/sieniawski1708/ · write-up

Not solved

Attempted, text not obtained; the notes say why the work stops.

Target	Date	Landed	Extent	Result	Where
Prince Maurice → Prince Rupert, Worcester, 7 July 1645 (Warburton iii. 133; Cryptiana item #5), and the royalist intercepts of May 1646, BL Add MS 72438 f. 9r (to "My Lord", 21 May, DECODE R8623) and f. 10r–v (to the King, 13 May, R8624; Cryptiana item #6); also f. 104r (R8725) and Arthur Trevor → Rupert, Bristol, [28] April 1645, Add MS 18982 ff. 51–52 (R8433)	1645–46	9 Oct 2026	—	Not deciphered; no key fits. f. 9 (152 groups), f. 10 (721), f. 104 (256) and Trevor's cipher runs (62) transcribed from the DECODE images; Maurice (93 groups) from the 1849 print. All 73 keys of Add MS 72438 (the Digby-coach keys, index f. 25r R8627, numbered 80–139 with their correspondents) screened by range; keys 118 (Ormond–Rupert), 123 (King–Queen), 125 (Mr Browne), Digby–Walsingham (R8724), R8744 and R8741 applied: no sense. The intercepts are of May 1646, seven months after the keys were taken; letter runs too short for a ciphertext-only rebuild, Trevor's 54 letter groups unstable under annealing. DECODE corrections queued (f. 104 is not "Decrypted"; R8433's author and date). What was deciphered beside them is f. 107 (row under Read with known key)	targets/rupert/ · write-up
Torcy → the French plenipotentiaries at Geertruidenberg (d'Huxelles, Polignac), Versailles 3 Apr 1710, and Villars → Polignac, camp of Arleux 1 Jun 1710 (copy); intercepted, BL Add MS 61575 ff. 38–41 and f. 44, DECODE R8755 and R8756	3 Apr and 1 Jun 1710	9 Oct 2026	—	Not deciphered; no key fits either code. Endorsed "Mr Blencowe cannot decypher them". Torcy: 690 groups (285 distinct, values 1–823), a two-part code of 800+ values with homophones; Villars: 173 groups (106 distinct, 3–560), a different code. DECODE transcriptions checked against the images (Villars 575 → 515); the clear opening of Torcy's letter, omitted on DECODE, transcribed. Keys tried: R8766 (Torcy–Petkum 1709), R8767, R8768, DECODE R587, Tomokiyo's 1710 Desmarets table (11B), BnF fr. 6204 keys: none fits. Legrelle iv 546 cites the French copy in AE CP Hollande t. 223 (not online); blocked on that copy.	targets/geertruidenberg/ · write-up
Military news for Prince Carl August Friedrich of Waldeck, unknown informant, Hessisches Staatsarchiv Marburg, HStAM 118 a Nr. 3954 ff. 3r–4r (HCPortal 517; catalogue 339, class C)	18 Jan 1744 (from the prince's draft reply on f. 4v)	28 Sept 2026	—	Not deciphered; the key is not identifiable from one letter. 717 three-digit groups (303 distinct) transcribed. The system is characterised as a two-part homophonic nomenclator: the writer struck a 9 and wrote 7 in a recurring phrase (7, 90, 924 are homophones); no alphabetical, column or multiplicative numbering fits; a synthetic R4680-style nomenclator reproduces the profile. On that control the true key scores far below gibberish keys, so ciphertext-only was closed. No key in Arcinsys, HCPortal or DECODE (112 Marburg keys of 1710–99 listed). The prince's French draft on f. 4v dates it and shows the writer seeking a place in his regiment. A key or deciphered copy may be in the undigitised 118 a Nr. 1973–1980 (Feldzug 1744). New lang model de-1740s.	targets/waldeck1744/ · write-up
Esterházy papers, box 636: six digit-cipher documents of 1744, one from Maria Theresa; the one page online is a despatch from Moscow, 5 Oct 1744, to "Votre Majesté", Slovenský národný archív, fond Esterházi – čeklíska vetva, box 636 (catalogue 344, class C)	1744	9 Oct 2026	—	Not deciphered; the box is not digitised and no key fits. Only Figure 6 of Antal et al., HistoCrypt 2023, is online: 176 groups (157 distinct), values 7–1737 plus 3018–3297, transcribed. The Vienna chancery keys of 1740–1746 on DECODE (R1562, R1574, R1593–R1598 incl. the 1745 French and German circulars issued to Rosenberg and Esterházy, R1601, R1644, R1691) do not reach those values. No reading found; a sibling (Maria Theresa to Esterházy, 31 July 1744, Dorotheum 17 Dec 2024) has its key in the ÖStA. Blocked on scans of box 636 and the 1744 key.	targets/esterhazy1744/ · write-up
Linear A, Minoan Crete: about 1,400 inscriptions (GORILA; texts via lineara.xyz and SigLA), a script with an unknown language, not a cipher	c. 1800–1450 BC	23 Sept 2026	—	Language not solved; the accounts give sense. Linear B sound values carry over (6 matches against 0.49 expected); 41.8% of tokens (numbers, fractions, commodities, KU-RO/KI-RO, three place names) give sense without the language; all 35 totals rechecked. HT 34 contradicts the published values of fraction signs H and A. Proposed Semitic and Luwian relatives fit no better than control languages. 67 rounds of tests describe the sound shape (rare e/o, mainly after R and T; consonant harmony), Minoan names at Knossos and the bookkeeping, until new tests stopped yielding results. No word translated.	targets/lineara/ · write-up
Indus script, Mohenjo-daro, Harappa and some 50 sites; Parpola, "Study of the Indus Script" (ICES Tokyo 2005) and Deciphering the Indus Script (1994); ICIT, M77 and CISI transcriptions	c. 2600–1900 BCE	23 Sept 2026	—	Not read; grammar set out. Parpola's structural claims hold; 7 + fish and fig + fish rest on one and two seals; 2 + fish, not 6 + fish, is the commonest count (also in Parpola's own CISI transcription and in M77). 2,364 hypotheses registered before testing (1,224 held, 1,140 failed): an ending slot chosen by the name's last sign; line breaks never split bound pairs or name + ending; numerals fixed with what they count; seal names an open, individual population like Linear B personal names at matched length (Ur III legends as a second control). No sound value	targets/indus/ · write-up
D'Agapeyeff challenge cipher	1939	15 Sept 2026	—	The ciphertext is not enciphered English	targets/dagapeyeff/
Vatican Challenge Part 5 (Farnese → Poggio): Cardinal Alessandro Farnese → Giovanni Poggio, nuncio in Spain, AAV Segr. Stato Spagna 1A ff. 70r–73v (DECODE R92; catalogue 262, with the open companion R91)	Apr 1542	20 Sept 2026	—	Solved earlier by Simon Klee, “The Farnese letter”, 16 Sept 2026; MysteryTwister accepted the submission. Found already solved. Klee recovered the same eighteen letter codes and null from each half independently, with synthetic and shuffled controls and checks against the manuscript. The diagnosis here after seven sessions is withdrawn: 4 is o, 9 is the null, and the core is a mixed one-/two-digit monoalphabetic key with dotted syllable and title codes, not an Antonio Elio polyphonic syllabary. 6,358 of 6,562 corrected digit tokens read exactly, 170 with recorded departures, 34 in gaps. Reviewed here; the image-verified transcript made here (151 digits corrected) stands. Open: R91 (Spagna IA-1), another key.	targets/vatican5/ · review · write-up
Dirk van Hogendorp (St Petersburg) → Maarten van der Goes, ciphered dispatch no. 12, Nationaal Archief 2.01.08 inv. 318 (DECODE R1942; catalogue 227, class C)	5 Jul 1803	20 Sept 2026	—	Solved earlier by Subject identified from Sillem 1890; exact plaintext open. Sillem’s account of Hogendorp’s failed approach to Chancellor Vorontsov says he sought another channel to Alexander I through General Jan Pieter van Suchtelen, and footnotes that paragraph to the ciphered dispatch of the same day, no. 12. The manuscript itself is headed no. 12. Authenticated DECODE access found the matching 1803 Russian-legation codebook, R1035 (1,849 entries in the partial CSV). Open: exact superposed marks and a group-by-group reading of all 286 groups.	targets/R1942/ · write-up
Rohan papers (d’Émery, Savoy, Canisy) and Louis XIII / Châtillon letters, BnF Français 5190 and 3758 (catalogue #320, class C)	1635–36	22 Sept 2026	—	Clear prose transcribed; cipher open. The Savoy and Canisy letters are printed deciphered, but their ciphertext and d’Émery’s letter lie beyond Français 5190’s Gallica scan, which ends at manuscript p. 80. Français 3758 nos. 100 and 104 were read wherever clear, including the king’s Valtellina force estimate and Châtillon’s account of Schenkenschans; their cipher blocks lack a verified key. Sainte-Geneviève MS 3338 is a second witness to seek.	research report · write-up
Niccolò Guidi di Bagno (Paris) → Secretariat of State (Rome), AAV Segr. Stato Francia 104 ff. 5r–7v (DECODE R5622)	5 Jan 1652	22 Sept 2026	—	All five images checked; four pages are an unseparated decimal cipher with Italian words left clear. Lasry's Francia 346 key is structurally excluded: it reserves 6 as a separator, while R5622 repeatedly writes 66 and 666. No DECODE transcript, attached document, clear copy, edition, online decifrati register or matching sibling was found. OCR and glyph clustering were too inaccurate to publish. Needs a manual double-pass transcription and the 1652 Paris nunciature key or Secretariat register	targets/bagno1652/ · write-up
Nevers → Villeroy, Saint-Quentin 16 Aug 1595, BnF fr. 3993 no. 102 ff. 148r–149r (file copy; catalogue item 277, class C)	1595	22 Sept 2026	—	(22 Sept 2026): 17 runs, 753 signs of figures and Greek-like signs, no decipherment. The key is not in Nevers's collection: fr. 3995 no. 76 (read in full) has the same sign family and design but other values; nos. 65, 66, 68–75 and the Balagny alphabet do not fit. A homophonic annealer that reads a matched control (753 letters, 49 signs + 3 nulls) letter for letter finds no French under six unit cuts, so it is a nomenclator with syllables/word codes whose marks the copyist dropped; the syllabic model's own control fails at this length. Needs the key or the as-sent letter in Villeroy's papers	targets/nevers1595/ · write-up
Baudouin-Desportes → Aldobrandini and → Frachetta, Paris 22 July 1593, BnF fr. 3984 ff. 186, 189 (catalogue item 20, class A)	1593	17 Sept 2026	—	(17 Sept 2026): the cipher is Tomokiyo's League polyphonic key (each glyph two letters), and the same-day letter to Lisieux (f. 188, deciphered on f. 184-185) gives 1,100 aligned glyphs of the hand; but a CNN trained on them reads only three glyphs in four on held-out lines, hand reading does no better, the polyphonic decoder needs over nine in ten, and the two letters share no verbatim passage with the Lisieux letter (alignment search against a shuffled control). c. 5,000 glyphs remain to be read; f. 177 (decipherment of f. 176) is the untapped training source	targets/sega1593/
Voynich manuscript (Beinecke MS 408)	c.1404–38	15 Sept 2026	—	Plain or simply enciphered European language excluded on transliteration-robust entropy; verbose encoding vs structured meaningless text left roughly even, with the separating tests named	targets/voynich/ · write-up
Pseudo-Elias of Cortona, Lumen luminum, North Italy c. 1525, Beinecke Mellon MS 29 ff. 1v–2v and front pastedown (DECODE R2877; catalogue 169 'Elia da Cortona to unknown recipient', C)	c. 1525	22 Sept 2026: all leaves checked; the 234-letter cipher (zodiac 'se chiama' table, Sol and Luna recipe lists) transcribed; monoalphabetic with a working control, homophonic, Caesar/Trithemius, Vigenère 2–7, Alberti ring with per-segment offsets, word-pattern solver	—	No key. Not a letter: an alchemical codex. Not simple substitution; IC 0.075 but no system tried gives Italian or Latin; possibly invented secret names (the Beinecke catalogue suspects charlatanism). Needs a clear copy of the prologue's recipes	targets/mellon29/ · write-up
Lodewijk van Toulon → Dirk van Hogendorp, The Hague → St Petersburg, 30 Jul 1803, NA 1.02.13 Legatie Rusland inv. 208 (DECODE R2034; catalogue 230, C)	1803	20–22 Sept 2026: 110 groups re-transcribed (10 corrections); the same-key 1801 Bourdeaux dispatches R1944 and R1946 transcribed in full and aligned with their solutions; R1035 (all mark mappings), R1038, R1891 and the other Batavian DECODE records tested	—	Key not digitised. Croiset's six-series Correspondentiecijffer, which survives only as NA 2.21.045/34313 (Museum B1 392). The 1801 solutions fix 3 of 67 words (394 handen, 58 brengen, 222= niet). Nine values claimed on 20–21 Sept were not reproducible and are withdrawn. The series are not alphabetical, so the rest needs the book.	targets/toulon1803/ · write-up
Gilbert Bayard → Montmorency, Toledo 5 Jan 1526, BnF Clairambault 325 f. 84 (17th-c. copy; DECODE R2285; catalogue 193, C)	1526	22 Sept 2026: 56 signs (29 kinds) transcribed; Champollion-Figeac 1847 CCXVIII prints the letter without the cipher; Calvimont (Biermann) and Raince 1526 keys tested; homophonic anneals	—	No key. Keys of the same months share the shapes, not the values; too short blind. Needs the original (V. 41 f. 4)	targets/bayard1526/ · write-up
Walsingham letter-book, Walsingham → Burghley and Elizabeth → Walsingham, Paris–London June 1571, BL Harley MS 260 ff. 112v–113r (DECODE R8356; catalogue 'Elizabeth R. to unknown recipient', C)	25 June 1571	22 Sept 2026	—	The Queen's letter is clear; the cipher is in Walsingham → Burghley, printed in Digges 1655 p. 106 with the cipher as figures. Nine-sign run 6174aH6HH and codes [12], [E4] unread: too short, no key	targets/r8356/ · write-up
Walsingham letter-book, Burghley/Leicester ↔ Walsingham, London–Paris Nov 1572, BL Harley MS 260 ff. 367–369 (DECODE R8361; catalogue 125, C)	7–27 Nov 1572	22 Sept 2026	—	'Unknown sender/recipient' = Walsingham's letter-book: four clear letters, printed in Digges 1655 pp. 297–301 with the cipher left as figures. [3] = the Queen by context (4 uses); codes [7], [9], H and a nine-sign letter run unread, no key known	targets/r8361/ · write-up
Benedetto Fantini (DECODE "Jantini") → the Este court, Buda and Eger, ASMo Ambasciatori Ungheria b. 4, Fantini nos. 5–6 (DECODE R1126, R1127; catalogue item 156, class B)	1517–1518	21–22 Sept 2026	—	Writer is Benedetto Fantini (Vestigia 1846–1852), not "Jantini"; no decipherment, clear copy or key in the file, on DECODE or in print. Two-tier sign cipher of the Caprile 1519 family, shown syllabic from Caprile R1130 against its 1882 decipherment 4c (upper = consonant, base = vowel, bracket = syllable). R1126's 1,370 columns transcribed; a syllabary annealer that recovers a same-size control (also at 10% noise) gives no Italian, and Caprile's values do not fit. Open: both letters; needs a key or a paired letter in Fantini's key	targets/jantini1517/ · write-up
Walsingham to Burghley, Paris, BL Harley MS 260 ff. 398-399, letter-book copy (DECODE R8363, catalogue #123, class C)	20 Jan 1572/3	2026-09-22	—	Clear English letter apart from one six-sign port name ("departed hence to 2 1 ⌘ ε 9 4 with intention to imbarke there") and a copyist's [9] placeholder. Digges printed it in 1655 with the cipher as cipher; CSP For. x no. 725 (same day) omits the passage. No 1572-73 Walsingham key; the Cecil–Norris 1568 table gives nonsense; the one other same-system group (R8364, "2 0 ⌘ H 6 1") leaves Dieppe, Calais, Nantes, etc. all possible.	targets/walsingham1572/ · write-up
Burghley to Walsingham, Westminster, BL Harley MS 260 f. 262, letter-book copy (DECODE R8358, catalogue #127, class C)	5 July 1572	2026-09-22	—	DECODE's 'unknown recipient' is Walsingham; clear English letter printed in Digges 1655 with the cipher as figures, CSP For. x silent. [3] = the French King by context and Walsingham's 13 July reply; [2], [4], [5], a monogram, A, F and the run 'm o R. d A.' unread, no key known	targets/r8358/ · write-up
Zeschau → Seebach, Dresden → St Petersburg, 18 Jan 1841 – 26 Oct 1843, HStAD 10731 Nr. 12 (DECODE R5005–R5008; catalogue, B)	1841–1843	2026-09-21	—	R5005 transcribed in full (70 lines, 3,969 unseparated digits). The rubbed-out pencil decipherment shows the cipher is a two-digit syllabary, not a homophonic letter cipher: 11 la, 70 pre, 82 m, 34 i, 29 er, 40 e, 46 que. No 1840s Saxon key on DECODE. Homophonic and syllabic annealers (control solves) give no French. Needs the key, UV imaging of the pencil, or R5006–R5008 transcribed	targets/zeschau1841/ · write-up
Anonymous ("MR" monogram) to the duc de Mercœur, BnF fr. 15564 f. 151 (catalogue #13, class C)	26 Jun [c. 1586]	2026-09-21	—	Six cipher runs in clear French, 217 signs of 55 kinds (Greek letters, □ △, digits). Not Lasry's 2022 fr. 15564 key (no shared signs; he had already judged it too short), not Mayenne–Forget or Clairambault 357; homophonic annealing scores no better than a shuffled control. Clear frame transcribed.	targets/mercoeur1586/ · write-up
Sir Henry Percy to Cecil, Norham, BL Add MS 4136 f. 172 (DECODE R9256; catalogue #87, class A)	23 Jul 1559	2026-09-21	—	The image is a Forbes tracing of the four cipher passages only: 63 signs in a box script with dot/underline marks. No single substitution fits them (pattern, constraint, Rosicrucian and annealing searches), no key in the volume; Cecil's decipherment is on the TNA SP 52/1 original, the calendars do not say which words were ciphered.	targets/percy1559/ · write-up
Paris 1719 and Carré 1742–45 code letters, BL Add MS 32305 (DECODE R2963–R2974; catalogue #78 and #70, class B)	7 Feb 1719 – 17 Nov 1749	2026-09-21	—	A codebreaker's file: R2963 and R2974 were deciphered at the time; the 1719 Paris letters (R2964/5) are a two-part 3-figure code with ~12 contemporary glosses (354 groups, 191 distinct in f. 6); the Carré letters (R2968/70/72, R2970 received 22 Aug 1742) were transcribed in full (1323 groups, 381 distinct) and tested: letter-homophonic anneal no better than a shuffled control, one-part order ruled out (r = −0.26), so a two-part code of ~800 entries with no gloss or crib; closed.	targets/bl32305/ · write-up
SP 53/16 nos. 78-79 and SP 53/22 f. 52 (Mary Queen of Scots papers residue)	1585	15 Sept 2026	—	Symbol ciphers of the Paris-Rheims exile network in Tomokiyo's glyph-label transcription, 507 and 644 groups. Second session: the two letters share one key (10 of the 20 commonest symbols in common, 2.9 expected), so they pool to 1151 tokens; third session: a numba pipeline now reads clean 507- and 1151-token controls (and a 507-token control with word-signs), yet no. 78 shows no language basin in English, French, Latin, Italian or Spanish, no. 79 sits at random level, and pooling scores worse than either letter, so the same-key claim is withdrawn; a nulls-plus-nomenclator control is not read at this length, so that design stays open. f. 52 (84 tokens) is below unicity. Needs the images and the SP 53/22 keys	targets/sp53/
Regent Moray → John Wood, Add MS 32091 f. 213 (134 groups, 32 symbols)	1568	15 Sept 2026	—	Homophonic alphabet in Scots on the face of it, and a 5-gram Scots annealer reads 5 of 6 matched 134-letter controls, but the target gives no reading in Scots, English, French or Latin; no null, separator or crib is supported. The catalogue says the passage is Border news, i.e. names, and 13% of genuine Scots windows score below the solver's false optimum, so the result is undetermined rather than excluded. Needs the page (BL offline) or a second letter in the cipher	targets/moray/
Louis XIV → duc de Chaulnes, Rome (300-group one-part code)	1690	15 Sept 2026	—	Ciphertext verified from the page images and the code's ten-column Croissy design established, but 300 groups do not determine a 116-entry nomenclator: the annealer recovers 4-12% of a matched control and wrong keys score within noise of the true one. No printed plaintext found. Needs the minute in AE Rome Corr. 331-332	targets/chaulnes/
Debosnys cryptograms	1882–83	15 Sept 2026	—	Cipher poem is rhyming couplets in a French syllabary, too short for any crib-free attack	targets/debosnys/ · write-up
Copenhagen cryptogram	c.1950s	15 Sept 2026	—	Two transcriptions, ten languages, six reading conventions; not a simple substitution of any language tested	targets/copenhagen/ · write-up
Scorpion letters S1 and S5	1991	15 Sept 2026	—	Below the unicity distance for a homophonic key; controls produce fluent false solutions. The 2018 claim tested	targets/scorpion/ · write-up
Charles I, Isle of Wight letters	1648	15 Sept 2026	—	Two letters still unread; two candidate keys newly excluded	targets/charlesi/
Berthier → Napoleon; letter to Marmont	1807–12	15 Sept 2026	—	Neither attackable from the single printed source (Vilcoq 1969)	targets/napoleon/
Kaliningrad bottle post	found 2015	15 Sept 2026	—	Blocker is transcription from two photographs, not cryptanalysis	targets/kaliningrad/
Koehler cryptograms (Abwehr)	1944	15 Sept 2026	—	Five short letter-cipher messages; skipped as intractable	targets/abwehr/
WW2 censorship-manual steganograms	1940s	15 Sept 2026	—	Blocked on image resolution; TNA's digital copy is the same scan	targets/censorship/
Lodovico Birago → Duke of Nevers, Saluzzo 13 Nov 1571 (BnF fr. 3251 f. 119)	1571	16 Sept 2026	—	Glyph-level transcription from Gallica: 483 digits, 16 marked digits, 9 inline wavy signs, 6 null letters. Variable-length prefix and suffix codes excluded against controls; the only consistent design is two-digit letters with one- or two-digit marked code groups (75 pairings, 228 tokens over 62 symbols), and at that homophony the annealer fails matched controls, so the target's negative is the method's limit; polyphonic single figures (Lucca-type) excluded with a method that reads its control 3/3; no sibling letter in the volume. Needs a sibling letter or a crib	targets/birago/
Colbert passages (Mélanges Colbert)	1665–75	15 Sept 2026	—	Read from the Gallica images: the Charost letter (Calais, 3 July 1675, not 1673) and the Gravel–Maulevrier copy (Mainz, 1674) are one key, Maulevrier's, 106 plain figures 52–489 between them. The printed catalogue located all 22 other Charost letters online and both Maulevrier letters: every one in clear. A monotone annealer reads matched controls of both strict alphabetical designs this office used and fails the target under each; looser designs fail their own controls. Probably a word nomenclator; needs the key or a clear copy of the news	targets/colbert/
Thurloe State Papers intercepts	1653–56	15 Sept 2026	—	Four short pieces, all period keys fail	targets/thurloe/
D'Estaing → Gérard	1779	15 Sept 2026	—	217 tokens of a 600-code with no key material; needs the archive copy	targets/destaing/
Le Tellier → Castelnau	1657	15 Sept 2026	—	Too short for an unconstrained syllabic solve	targets/letellier/
Cardinal de Joyeuse → Villars, Rome, 15 Feb 1594 (500 Colbert 33 f. 539)	1594	16 Sept 2026	—	Transcribed from Gallica: 358 tokens in three layers (81-sign homophonic alphabet, 48-sign marked syllabary, 29 code numbers); not Lasry's Senecey key from the same volume; the letter-layer anneal recovers 2-18 % of three matched controls and the target's seeds disagree	targets/joyeuse/
Cocquet → Mangot, Rome, Nov 1616 (Clairambault 369 f. 317)	1616	16 Sept 2026	—	Gate check: ~160 letter-shaped glyphs in nine runs inside clear French; none of the five period keys from the same volumes matches; du Croc regime	targets/cocquet/
Louis de Gonzague, duc de Nevers → Robert de La Vieuville, Saint-Aignan, BnF fr. 3975 item 34 ff. 101–102 (DECODE R4289), catalogue item 184, class C	30 Sept 1587	22 Sept 2026	—	A copy of a clear letter in a very hasty hand with code numbers and three short runs set in; the BnF notice's "déchiffrement" is not on the leaves. Tomokiyo's Nevers key no. 16 (fr. 3995 f. 32v, transcribed) fits the code numbers (42 et 43 eschevins et maire de ville, 63̄ La Vieuville, 48 château, 30 artillerie; ~31 items graded in items.tsv) and a phrase of the 31 Aug 1587 sibling, but the runs (33 signs) give no French under any reading it allows. Letter text illegible on the B/W microfilm (~30–40%); a noise control shows a solver needs ~80%. Not in Gomberville's Mémoires (1665). Blocked on a colour scan or the original	targets/vieuville1587/ · write-up
German Ordnungspolizei Doppelkasten radiograms, Mogilev/Rovno/Proskurov, 27 Feb and 16 June 1942 (Weierud 2020, NARA RG 457 HCC Boxes 1386 and 202; Schmeh, Cipherbrain 30 Aug 2020)	1942	18 Sept 2026	—	Double Playfair, each pair enciphered twice through the same two boxes. The discriminators show that the seven parts of 27 February (970 letters) are one daily key. Single encipherment is excluded; for the double pass, annealing, tempering, staged scoring, Sinkhorn relaxation and EM all fail on a synthetic 970-letter control with a known key (true-key basin 8-10 swaps; fails even with the cell permutation given). A known-plaintext solver recovers the boxes from ~80 pairs, but Bletchley's decrypt of the 16 June message (Borki) does not align with its ciphertext under any spelling, layout or length tried. Needs the February decrypts (TNA HW 16/17) or the exact enciphered wording of the June message	targets/orpo1942/ · write-up
Catherine de Médicis → Philibert du Croc, Saint-Maur, 27 Apr 1567 (Destray 1924, facsimile plate)	1567	16 Sept 2026	—	Transcribed from the Gallica plate; a symbol cipher too short for an unconstrained solve unless the ticks turn out to be separators	targets/ducroc/
Queen María of Castile → Alfonso V, Valencia, 5 Aug 1435 (day of Ponza), ACA Cancillería reg. 3225 f. 59r (DECODE R10171, public, "non-decrypted"; oldest-cipher candidate A1)	1435	18 Sept 2026	—	42 signs in three runs inside clear Catalan; the clear context points to the truce of Majano ("es continuada daci a tots sants"). The 1437 Montcada–Amat key (DECODE R10183) is a different alphabet; the 1429 Castile-war key (Cortès & Pons, Saitabi 36, 1986) does not fit as an alphabet, but its nomenclator seems to match the dotted word signs (regne d'Aragó, Catalunya, rei de Castella, Navarra): a hypothesis, not a reading. Regs 3226, 3224, 3113 and the king's 2693 swept in full on PARES (1,414 images): no cipher anywhere; the king's register shows his letter to the queen of 27 July 1435 sent "in ciffra". Crib-anchored annealing with a Catalan model gives nothing. Needs reg. 3225's other 1435 folios (ACA, not digitised), the second key of B.A.H.M. ms. 677 ff. 160–161, or a queen-side key at the ARV	targets/maria1435/ · write-up
Stepney → Manchester, Vienna	1702		—	Stepney's office cipher in TNA SP 105/106 or BL Add MSS 7058–78	targets/stepney/

Not applicable

Not a cipher, or the item could not be reached.

Target	Date	Landed	Extent	Result	Where
Stojan (a Serb prisoner) → aghas, knezes and his Vlach kin, ÖStA HHStA Ung. Akt. Spec. Verschwörerakten VII, Fasc. 327 Konv. D (Chiffres 1664-1668) ff. 30-31 (DECODE R688, catalogued as 'Unknown sender to unknown recipient, 1 Jan 1664', simple substitution in graphic signs; catalogue 270, class A)	undated (bundle 1664-68)	22 Sept 2026	—	Not a cipher. The graphic signs are Cyrillic letters writing vernacular Serbian: Stojan greets aghas, knezes and named Vlachs and begs them to get him out of the German prison (извадить из немачкий тамнице); signed as written for him by Mla[den?] Baki[ć?]. Slip on f. 31 = second damaged note. Transcribed line for line, about a quarter of the words doubtful; place names Z?vač and Bor? open.	targets/r688/ · write-up
Walsingham letter-book, Burghley → Walsingham, Westminster–Paris, BL Harley MS 260 f. 395 (DECODE R8362, catalogued as 'W. Baurleighs? to unknown recipient, 27 Jan 1572'; C)	24 Jan 1572/3	22 Sept 2026	—	Not a cipher: a clear letter, read in full. Letter-book copy of Burghley's letter, printed in Digges 1655; DECODE's '27' is '24th'. The only signs are boxed 9s after the cover-names Lyons and Roome, which mark them as cover-names. Who they stand for is not known	targets/r8362/ · write-up
C. Ben. Schaeblin (The Hague) to Mr Jones, BL Add MS 32256 f. 235 (DECODE R9218, catalogued as "Mr. Jones (Hague) to Ben. Schaeblin", nomenclator, numerical, non-decrypted; catalogue 72, class A)	16 Jan 1753	21 Sept 2026	—	Not a cipher: a clear English letter, read in full. Sender and receiver are swapped on DECODE. Schaeblin proposes a new French cipher of 3000 numbers with blanks, modelled on the English one, and disguises by reversing or rotating the figures of each group or adding a dummy figure, each marked by a "Change the Cypher" group; the only numbers are worked examples on 1753. The French cipher itself not looked for	targets/jones1753/ · write-up
Charles V → Prince Philip, minute on an exchange with Anton Fugger, AGS Estado leg. 1563 fol. 572 (DECODE R9970, catalogued after Galende 1994 as "Andrea del Burgo (Ferrara) to Chancellor Gattinara, 26 Oct 1527", alphabet, graphic signs, numerical, non-decrypted; catalogue 198, class C)	1543–1554	28 Sept 2026	—	Not a cipher: two drafts of a clear Spanish minute "al príncipe". The Emperor tells Prince Philip of a further exchange of 110,770 ducats ("CX U DCC LXX") with "Antonio Fucar", principal and interest to be paid at term. Philip was born in May 1527, so not 1527. The Roman numerals and U thousands sign account for DECODE's tags. The Burgo letter of 26 Oct 1527 is not on the record and was not found on PARES, in CSP Spain III.2 or in Tomokiyo	targets/burgo1527/ · write-up
Count Visconti (Brussels) with Marquis Visconti (London), BL Add MS 32270 f. 42 (DECODE R7931, "Count Visconti to Visconti", "partially decrypted" ciphertext; catalogue 73, class A)	1727	21 Sept 2026	—	Not a letter: the key. The Deciphering Branch's fair copy of the Italian nomenclator it rebuilt (1–300, 193 filled: letters with homophones, doubled letters, CV syllables, names from 1 Walpole, 22 Sinzendorff, 146 Ostenda, 170 Townshend to 254 Staremberg), endorsed "Count Visconti at Bruxelles with Marq.s Visconti here, 1727, Italian Cypher"; R7932 is its working draft, R7933 a French key for the same pair. Nothing enciphered on the leaf; no letter in this cipher in the volume	targets/visconti1727/ · write-up
Mr Eichel → unknown recipient, "Prussia 1758", TNA SP 106/7 image 0020 (DECODE R595, "Non-decrypted", "probably from Russia"; catalogue 65, class A)	1758	21 Sept 2026	—	No message: nulls and blank key cells only. DECODE's key R594 does not fit. Key R596 (same volume) declares every number with a 7 and 1–503, 1115–1198, 1483–1999, 2600–2802 null: 117/139 groups are nulls, and 21 of the other 22 are cells R596 leaves blank (7% of live numbers). A dummy sheet; R596 undated, its tie to Eichel open	targets/eichel1758/ · write-up
Instructions for the cipher with Baron Stosch and Mr Walton, TNA SP 106/7 f. 33 (DECODE R597, "unknown sender to unknown recipient", "partially decrypted", "lines unsolved"; catalogue 64, class B)	c. 1731	21 Sept 2026	—	Not a letter: an instruction sheet, read in full. Two keys (A black, B red) of three-digit syllable groups 100–999, 371 = change of key, groups run on unseparated, nulls 110, 220 &c. and all above 900. The worked example Les Adherens du Pretendant se donnent beaucoup de mouvement à Rome has its syllables written over every group; DECODE's unsolved lines are the same example run together. Stosch = "John Walton", Walpole's spy on the Stuart court; Rome corrected to Florence dates it c. 1731. 702 (du/up) and 023 doubtful; keys A/B not on the record	targets/sp106r597/ · write-up
W. Loeschner → unknown, Vienna, ÖStA HHStA Staatskanzlei Interiora, Chiffrenschlüssel Kt. 20 ff. 30–40, DECODE R2227 (catalogue, 'W. Loeschnen?')	1 Jul 1816	21 Sept 2026	—	Not a cipher. A clear German memorandum, Zusätze und Abänderungen, proposing changes to the instruction for using the Staatskanzlei's cipher keys, signed W. Loeschner. Its figures are worked examples with the plaintext written above them (le Roi d'Espagne vient de nommer … M. le Duc de San Carlos). Nothing hidden	targets/loeschnen1816/ · write-up
Unknown sender → unknown recipient, BL Sloane MS 3188 ff. 109–169, DECODE R8506–R8530 (catalogue 138 + 139, class C)	17th c. (leaves dated 1594–95)	21 Sept 2026	—	Not a cipher. The 25 leaves (DECODE R8506–R8530) after John Dee's spirit diaries are notes attributed to William Shippen, in 17th-century English stroke shorthand with names, titles and years in longhand: an index of Dee's Actions, a year list of his writings, a note on Kelley, Pucci and Prague (1595). Shorthand not identified or read	targets/sloane3188/ · write-up
Frederic, Re di Prussia → unknown recipient, BL Add MS 61575 f. 95, DECODE R8774 (catalogue 98, class A)	1702–23 (volume span)	21 Sept 2026	—	Not a cipher. No numbers on the page: seven lines of plain French, spelled by ear and split into syllables, thanking the recipient for teaching the writer d'écrire en chiffre and wishing them a good journey. Writer and recipient unidentified	targets/prussia1702/ · write-up
Hyde's ciphered superscriptions	1659–60	15 Sept 2026	—	Dummy numbers "only to puzzle the Enemy", per the 1724 editor and the full Hyde–Barwick key of 1721	targets/hyde/ · write-up
Chinese gold bar cryptograms, Shanghai	1933	15 Sept 2026	—	Almost exactly ten of every letter; flatter than any cipher of a real text can be. No message	targets/goldbar/ · write-up
Roosevelt cryptogram, number block	1935	16 Sept 2026	—	A permutation of 1 to 52, each once, padded with zeros; its statistics are those of a list written by hand, and the testable ordered-key cipher readings fail while matched controls succeed. Ernst's 2017 doodle claim confirmed. No message	targets/roosevelt/ · write-up
Beale Paper no. 1	1885	14 Sept 2026	—	Fabrication; evidence in notes, book-cipher scan over Gutenberg negative	targets/beale/
Intercepted League and Spanish letters summarised for the duc de Nevers, Sept–Oct 1589, BnF fr. 3977 no. 96 (catalogue item 31; "whether the intercepts survive in the volume is unknown")	1589	20 Sept 2026	—	They do not survive, and the summary is not in cipher. No. 96 = f. 318 = canvas 520 is a digest in clear French, fifteen numbered points: Mercœur asking Spain for help, powder and money carried to Nantes by Diego Maldonado "avec le chiffre général d'Espagne", Mauléon and "la lettre escripte du chiffre qu'ilz ont ensemble", Mendoza at Bayonne on Arques and Dieppe, Spain pressing the Pope to renew the excommunication. Every piece in the volume dated September or October 1589 (nos. 81, 84, 85, 89–97) is in clear, so there is no ciphertext to attack under no. 96. What the sweep found instead: eleven other intercepted ciphered letters in the same volume, Feb–Aug 1589, each with its contemporary decipherment written word by word above the cipher — nos. 29, 31, 38, 39, 43–45 (three from Mendoza to the duke of Parma, filed in a multi-folio entry a naive regex drops), 47, 48, 49, 71. Eleven complete cipher-and-plaintext pairs for Spanish ciphers of 1589; none transcribed, no key rebuilt	targets/nevers1589/ · write-up
Zhongshan telegrams	c.1938	15 Sept 2026	—	No corpus exists online; the premise of the list entry was wrong	targets/zhongshan/
Blancmesnil → Nevers (fr. 3633 f. 24)	1589?	16 Sept 2026	—	No ciphertext online: fr. 3633 is not digitised (the source page's catalogue link goes to fr. 4736); the Potier-Nevers key of 1589 is the candidate once a copy exists; sibling letters with decipherments sit in fr. 3616/3621 (DECODE R9434-9448, images restricted)	targets/blancmesnil/
Charles II → Duke of Hamilton	1650		—	Copy order for NRS GD406/1/2197 (open)	targets/hamilton/

Five further list entries turned out to be solved by others with nothing to add here: Perwich 1670 (Brown; Lasry, Biermann and Tomokiyo, 2025), the Milroy telegrams (Bean, 2026), the Feynman ciphers 2 and 3 (2023), the Confederate Navy dictionary code (2026) and the Mazarin–Bordeaux letter of 1654 (Lasry, 2025). They are not counted in the results above; the directories targets/perwich/, targets/milroy/, targets/feynman/ and targets/barney/ hold only the pointer to the published solution.

In progress

Target	Date	State	Where
“Mr. Conleys Hand” (Abraham Cowley's hand), Paris → a royalist in Holland, 28 Dec 1652 and 3 May 1653, BL Harley MS 7003 ff. 314–315 (DECODE R7768–R7769; catalogue 137, C)	1652–53	Clear text read; code groups unread (21 Sept 2026). The docket reads "mr Cowleys Hand", not "Conleys". Two clear English letters from the Queen's court with 23 code numbers in five runs (names, a place; 248:88:46:47:96:52:7:57 perhaps "Scotland", M). Too few for ciphertext-only; Hyde–Barwick key ruled out; not in Miscellanea Aulica. Needs a Jermyn/Queen key of 1652–53	targets/conley1652/ · write-up
Van Dedem → Van de Spiegel / Van der Goes, Constantinople 1788–1799, NA 3.01.26 inv. 212 and 2.01.08 inv. 346 (DECODE R2120, R2122, R2131, R1895; catalogue 225, C)	1788–1799	Unsolved; key not found (21 Sept 2026). R2120 was read at the time (decipherment filed as R2121, 22 Dec 1788, the Ottoman loan). R2122 (258 groups) and R2131 (50) use the same homophonic word code; no key on DECODE, and cribs R2121/R1947/R2053 cover only ~55% of the groups and do not align to a consistent key. R1895 (1799) uses a different marked system	targets/dedem1788/ · write-up
Palm → Starhemberg / Visconti, London 1727, BL Add MS 32270 ff. 16–18, 41 (DECODE R7927, R7930; catalogue 74, A)	1727	Unsolved; key not found (21 Sept 2026). R7927 = German letter, mostly clear, 14 homophonic+nomenclator passages (162 letter tokens, 47 values; glosses 33=e 23=n 10=d 15=t 40=ſ 230=der); R7930 = cover + French/Italian worksheet only, no ciphertext imaged. Windischgrätz key R7928 does not fit; 5-gram annealing failed. Needs Palm's key or the clear copy	targets/palm1727/ · write-up
Van Spaen → Van der Goes, Düsseldorf 14–15 Jan 1808, NA 2.01.08 inv. 281 (DECODE R1941; catalogue 226, B)	1808	Unsolved; key not found (21 Sept 2026). Letter (228 groups) and annex (75) of a plain numerical code to at least 1339, transcribed. Inventory: inv. 281 = ceded-districts commissioners (Zevenaar/Berg), not a Westphalian embassy. No DECODE key; Croiset R1035 ruled out; not in Colenbrander V. Needs the annex reading or the commission's code	targets/spaen1808/ · write-up
Unknown sender → unknown recipient, 1728, BL Add MS 32287 ff. 37–40 (DECODE R7707; catalogue 77, B)	1728	22 Sept 2026	Attempted, not read. One page of 205 invented cursive signs; the other leaves are a contemporary codebreaker's tallies and contact tables (no reading) and a docket 4 Aout 1728. IoC 1.07: not a simple substitution; mono, Vigenère and Quagmire-type anneals failed. Needs the key or the sibling letters
Sforza side → Giorgio (Zorzo) del Maino, reply, 4 May 1446, BnF italien 1583 f. 68 (DECODE R7898; catalogue 190, C)	1446	Not read; key not found (21 Sept 2026). Identified from Mazzatinti 1883 ("Risposta a quanto ha riferito Zorzo Maino… In cifre"); clear Lombard Italian with graphic-sign runs, 404 signs transcribed, plus the same-day f. 70 (Vincenzo Amidani, found on R7899, 390 signs). Annealing fails; the related Simonetta–Vincentius key (ASMi Sforzesco 1597 no. 14) shows syllable and word signs, but no DECODE Sforza key matches. Needs the 1446 Maino key or a colour scan	targets/it1583/ · write-up
Roell (?) → Van Dedem, Amsterdam → Constantinople, 9 Feb 1809, NA 1.02.04 legatie Turkije inv. 804 (DECODE R1469, R1470; catalogue 233, C)	1809	Unsolved; key not found (21 Sept 2026). Two unsigned letters, 13 pages, 2,585 groups of a heavily homophonic code to ~3,300 (no group above 1%). No decipherment on the scans; not the Dedem 1788–93, Spaen 1808 or Fagel 1804 code; no key on DECODE or in Roell's papers (which hold Croiset letters 1808–09).	targets/roell1809/ · write-up
Robert Fagel → William V, The Hague 18 June 1804, KHA Prins Willem V old inv. 337 (DECODE R2238; catalogue 219, C)	1804	Unsolved; key not found (19–21 Sept 2026). 128 groups of a French numerical nomenclator to at least 2510, transcribed and placed in the 1804 Orange compensation talks. Grand Chiffre R1024 (direct, additive table, 5,000 renumberings), all KHA and Fagel keys in DECODE, BL and Uppsala French codes ruled out. Key most likely in undigitised KHA A31-902; archive request drafted	targets/fagel1804/ · write-up
Hellen → Frederick II, KHA Prins Willem V inv. 196 (seven catalogue targets plus R1046)	1752–1763	Unsolved; investigation open (20 Sept 2026). Transcription parser audited; repeated passage confirmed on the manuscript. Fagel 5206 lists deciphered letters from 1752–1753, but no matching text or key has been verified.	targets/hellen1752/ · write-up
"BLUME SALAMANCA" telegrams, Zurich → London	1937	Transposition of telegraphic Spanish. Every single transposition and every double columnar with a second key of ten or fewer letters excluded against planted controls (lag scan plus exhaustive second-key enumeration). Open: double columnar with two long keys; needs the second telegram. Transcription corrected in three letters (L misread as I) on 30 Sept 2026 after Richard Bean's issue 14; the exclusions stand. Attempted, not solved	targets/blume/
"KR Blitz" Enigma message, OB Oberrhein → OKW, 10 Jan 1945 (Schmeh 2019)	1945	Ciphertext-only attack begun 16 Sept 2026; the first pass is partial and has produced no convincing plaintext	targets/enigma/

Surveys

• research/cryptoole/ — open records from Satoru's Cryptoole index (DECODE, Cryptiana, Rosson) not yet in the catalogue, matched against the repository by R-number; also flags catalogue entries Cryptoole now lists as solved.
• research/top50/ — Schmeh's Top 50 cross-referenced entry by entry and re-checked against what has been solved since each post.
• Thirty favourites — ten funny, ten historically significant and ten ingenious findings from the write-ups, each linked to its page.
• George Lasry's BnF cipher research — 56 enciphered letters in the BnF (1452–1859), one write-up page each deciphered by George Lasry with ChatGPT Astra 6 and Anthropic Opus 5.5 (10 Oct 2026), each spot-checked here against Gallica; 55 filed as targets/<slug>/ with Status: no write-up.
• Why the famous ciphers resist — Kryptos, Voynich, Dorabella, Beale, Linear A, Phaistos, the pigeon message, sorted by the actual reason each holds out.

Repository layout

README.md          this file
CLAUDE.md          working rules for the AI sessions (write-up, profile, shared-checkout conventions)
catalogue.json     the open targets, with scores and status; CATALOGUE.md is generated from it
CATALOGUE.md       the archival catalogue: how it was compiled, the open table, what was removed and why
SOLVED_CATALOGUE.md  what was solved or read, how, and where the AI did the work
SOLVED_RANKING.md  the solved targets ranked by difficulty and historical weight
profile.schema.json  the schema of every targets/<name>/profile.json
unsolved.htm       snapshot of Tomokiyo's source page, for diffing against later revisions

.claude/           session skills (/writeup, /profile) and the hooks that run the write-up audit
docs/              the website (GitHub Pages) and its build and check scripts, see below
lang/              shared language models: corpus registry, model registry, one n-gram engine (lang/README.md)
papers/            publication drafts;  unpublished/  pages held back from the site
decode_updates/    corrections and transcriptions queued for the DECODE database
tools/             shared scripts: transcription export (transcription_xml.py), the targets/ migration for old checkouts

targets/<name>/    one directory per cipher target: NOTES.md, profile.json, ciphertext, scripts, small derived data
research/          work that is not a single target:
  catalogue_harvest/  bulk DECODE / archive harvests that feed catalogue.json
  cryptoole/          catalogue candidates pulled from Satoru's Cryptoole index (pull.py)
  gallica_sweep/      Gallica volume sweeps for unlisted cipher leaves;  gallica_siblings/  their sibling letters
  oldest/             the search for pre-1497 ciphers (CANDIDATES.md)
  top50/              Schmeh's Top 50, cross-referenced
  mtc3/               MysteryTwister challenges, the harder track

The skills and checkers take a target's bare folder name (/writeup gramont1529, python docs/_check_profile.py gramont1529) and look it up under targets/. A target script that puts .. on sys.path still finds the shared lang package through the targets/lang/ stand-in; new scripts should use ../... Anything a second target could reuse belongs in a shared directory: language models in lang/, not a new targets/<name>/lm.py.

Every target directory has a NOTES.md except targets/barney/ (a pointer to a published solution) and targets/richelieu/ (its record is SOLUTION.md). Large downloads (microfilm, corpora, OCR, run logs, images) are excluded by .gitignore and regenerated by the scripts named in each directory's notes.

The website

docs/ is served by GitHub Pages. One manifest in docs/_build_site.py drives the navigation, footers with previous/next links, "On this page" strips and the index cards. After editing any page:

cd docs && python _build_site.py

The builder is idempotent. It also regenerates the "At a glance" counts at the top of this file. The priority queue on the index page is generated separately from docs/_queue.json by python _build_queue.py, and the catalogue page from catalogue.json by _catalogue_page.py.

Two checkers keep the ledgers in step. python docs/_check_writeup.py <slug> lists every surface a finished target must reach (page, manifest, README row, ledgers, catalogue, DECODE queue) and prints result: complete when it has; --audit lists finished targets with no write-up and the drift between README, manifest and pages. python docs/_check_profile.py <folder> validates a target's profile.json against the schema.

Credits are data, not prose. Each profile's outcome.credits names who contributed what (ids in contributors.json; roles in profile.schema.json), and outcome.provenance says whether the result is the project's, contributed or mixed. docs/_credits.py reads them; the build prints a credits line in each write-up's hero, generates docs/contributors.html (docs/_contributors_page.py: the table by person and kind, a card per contributor, the scholarship the work builds on, the queue of readings waiting for an independent check) and the front page's top-contributors strip, and writeups.html#by=<id> filters the write-ups to one contributor. docs/contribute.html and CONTRIBUTING.md are hand-written; the issue forms are in .github/ISSUE_TEMPLATE/. Credits for work before 13 Oct 2026 were backfilled by tools/backfill_credits.py from the key web, page bylines, merged pull requests and issues, and the notes.

The sender and recipient portraits under a write-up's title come from docs/_portraits.json (slug → role, name, image, what the picture is, and its Wikimedia Commons file). Only public-domain portraits are used, cropped to a 192-px square as docs/portrait_<name>.jpg. Leave a side out when the correspondent is anonymous, an office, or has no trustworthy portrait.

Datestamps are kept in docs/_dates.json: for each page, the day the finding first landed in the repository and the day the page last really changed. _build_site.py maintains the second by hashing the page with everything it generates itself removed, so a rebuild that only restamps the furniture does not move the date; a new page, or a new line in "Recent findings", dates itself on the next build. The file was seeded from the git history by python _seed_dates.py, and any date can be corrected by editing it (the seeder never overwrites a date already there). python _stamp_markdown.py writes the same dates into the result tables of this file and of SOLVED_CATALOGUE.md.

Reproducing the two flagship results

Armstrong → Madison, 30 August 1808

John Armstrong (U.S. Minister to France) ended a private letter to Madison with 49 groups of the diplomatic code he used 1804–1810, in which 972 = the. The code was never published. It was reconstructed from the State Department's own pencil interlinear decodes on NARA microfilm M34 roll 13 (frames 0192–0201, a despatch of October 1806), merged with Tomokiyo's known-plaintext values from the letter of 4 May 1806, and completed by alphabetical-slot inference. Reading:

Russel ought to be the consul: he is an American by birth, and is much better qualified than any other candidate. In a word, he is above men in general. Next to him in fitness is O'Mealy, but he is, like Warden, an Irishman.

file	purpose
targets/armstrong/NOTES.md	group-by-group evidence, corrections to the Founders transcription, residual doubts
targets/armstrong/pairs.txt	~500 number → syllable pairs read from the pencil decodes, with frame/line reference and H/M grade
targets/armstrong/code972_partial.json	Tomokiyo's table from the 4 May 1806 known plaintext (base layer)
targets/armstrong/decode972.py	merges the three layers and renders any coded passage: python decode972.py ps|feb|all|table
targets/armstrong/export_key.py → key972.js	exports the merged 580-group table for the website decoder
targets/armstrong/pencil_score.py	ranks microfilm frames by amount of faint pencil (finds the annotated despatches)
targets/armstrong/crawl_wb.py	follows Founders Online correspondent links via the Wayback Machine to list coded letters

Not in the repo: the 393 roll-13 frames (2.5 GB, from NARA catalog 188671172), roll 14, crops, and the downloaded Founders/LOC pages. The decode reproduces from tracked files alone:

cd armstrong && python decode972.py all

Richelieu → M. de Rancé, July 1629

Homophonic substitution recovered by ciphertext-only analysis from a published transcription, then found to agree word for word with the decipherment printed by Avenel in 1858 (Lettres … du cardinal de Richelieu, t. III, pp. 368–369, 381–383). The letters are listed as undeciphered in current catalogues (cryptiana; DECODE R9461–R9462) and should be marked solved. Full write-up: targets/richelieu/SOLUTION.md.

file	purpose
targets/richelieu/richelieu1629.txt	S. Tomokiyo's transcription (source)
targets/richelieu/parse.py	tokenise transcription, frequency / n-gram stats
targets/richelieu/build_ngrams.py	build French quadgram model from Gutenberg texts
targets/richelieu/solve.py	simulated-annealing homophonic solver (--fix, --core, --weight)
targets/richelieu/render.py, final.py	render letters with a hand-built / final key
targets/richelieu/avenel.py, avenel_ctx.py	fetch and search Avenel vol. III OCR (Internet Archive)

cd richelieu && pip install requests && python build_ngrams.py && python solve.py --restarts 8 && python final.py

Conventions

• Priorities (George Lasry's advice, 24 Sept 2026). Time goes first to original solutions, attempts to recover a key from the ciphertext alone with the strongest model available; second to solutions that match the ciphertext with a key or plaintext found elsewhere (a printed edition, a calendar, a key filed in another collection). Targets whose work would only be transcribing and applying a key already identified, or transcribing an existing decipherment, come last; they still get a DECODE correction when one is due. Open targets are in catalogue.json; harder problems outside the historical corpus are in research/mtc3/.
• Every claimed reading is graded: H read from a primary key source, C from a known-plaintext letter, M uncertain, I inferred from context or alphabetical position. The website decoders show the grade per group.
• A negative result is only reported alongside a matched control: a synthetic text of the same length, alphabet and cipher design that the same solver does recover.
• Before treating a catalogue item as unsolved, check the 19th-century printed editions (Avenel, Camden Society, Nuntiaturberichte) and the comment threads of the list posts. Six items so far were already solved in the open.
• Extent: complete or partial. A target's text is complete when (1) at least 95% of its cipher tokens give sense (fraction_coherent where set, else fraction_read, measured from a reading file, not estimated); (2) no gap is left untried; (3) every document in it is covered (judged per letter: a group with one letter left open is partial); and (4) what stays open is scattered code groups or names, or pieces blocked from outside (no key material, too short, illegible, needs the archive). Otherwise it is partial. (In profile.json the extent is outcome.class: read / read in part.) The state of the key is recorded separately (outcome.key: recovered / partial / none), and open nomenclator codes are counted apart (outcome.codes_open), because historians call a letter deciphered with scattered code gaps. There is no field standard to borrow: DECODE's status labels are assigned by each record's owner (Megyesi et al., HistoCrypt 2019), and computational work calls ~95–98% token accuracy "solved" without fixing a cut-off (Ravi & Knight 2011: Zodiac-408 at 97.8%). The bar, and the no-untried-gap condition, are this project's own.
• Outcome method. Every target is classed by how its text was obtained, in George Lasry's categories (email of 21 Sept 2026) plus one for complete existing decipherments: key recovered from ciphertext-only; key recovered based on plaintext from external sources (an edition, a calendar, a copy elsewhere); key recovered based on adjacent plaintext (a partial interlinear or marginal decipherment, clear passages, a deciphered copy in the same file); read after matching with key from external sources (an existing key not tied to the document, found here to fit); read with known key (a key already identified for the document or its correspondence); read from existing decipherment; not solved; not applicable (not a cipher, or not reachable). The first three are cryptanalysis done here. The method is outcome.method in profile.json, with a sentence of evidence (method_basis) and what the work added (contribution: transcription, decipherment with an ambiguous key, key recovered, key extended or corrected, sender or date identified, catalogue correction). Definitions: docs/_methods.py and the glossary. The Results tables, the site badges, the write-up filters and the scoreboard follow the profile; python docs/_check_writeup.py --audit section G lists any disagreement. Whether and when an earlier solution was found is recorded apart (prior_solution): a target whose key was recovered here before an earlier solution came to light is an independent re-solution and is marked as such. A target read by more than one route is classed by its primary achievement (Lasry, 25 Sept 2026); when the achievements are of genuinely different types, the entry is split into parts (outcome.parts), each with its own method, extent and Results row, and outcome.method names the primary one.
• Words. "Read" and "solved" are not used alone as outcome words, because they mix method and extent. The pages say what happened (deciphered with a known key, key recovered from the clear passages, the contemporary decipherment transcribed). A catalogue's "undeciphered" (Tomokiyo) means no contemporary decipherment is known, not unsolved. python docs/_check_terms.py <slug> lints a page for these and for the style tells.
• Dates in notes are absolute. Sessions are dated so that "since" claims can be checked against the source lists' last-modified dates.

Publication drafts

papers/ holds unvalidated drafts, none submitted: histocrypt/, anonymised HistoCrypt-format papers on individual results with the official style files and a shared bibliography; cryptologia/, a Cryptologia version of the Armstrong paper; and lasry/, the material for the paper with George Lasry on how LLMs perform against historical ciphers, built from the targets' profile.json records. papers/README.md has the format rules and the pre-submission checklist.

Contact

Daniel Bourdeau, [email protected]. Corrections, prior solutions I have missed, archive copies, or pointers to key material are all welcome. Issues and pull requests on this repository work too.

Licence

Text and notes CC BY 4.0; code MIT. Manuscript images come from many holders (among them the BnF, the British Library, The National Archives, the Library of Congress, NARA, the Vatican and Venetian archives through DECODE, the Hessian and Bavarian state archives, JACAR and the IACR) and remain subject to those institutions' terms. Every image on the site is credited under it to its specific source (holder, shelfmark and folio, and the route: Gallica, DECODE record, IIIF, edition); crops of BnF cipher photographs are used with attribution.