Paste a job ad and see your real chance of getting past the resume screen — before you spend an hour applying. Then see exactly what you're missing, get your resume and LinkedIn wording improved for that job, and track every application. Free, private, no account.
- Job seekers who send lots of applications and want to spend time on the ones they can actually get.
- Career switchers who want to know which requirements they already meet and which ones are real gaps.
- Anyone tired of ATS keyword scores that say "72% match" without saying whether you'll get a call.
| Feature | What you get |
|---|---|
| Resume screen chance | Your estimated chance of getting past the first resume check (recruiter or ATS), with an Apply / Worth a try / Probably skip verdict and plain reasons. |
| Requirement checklist | Each must-have from the job ad marked ✓ yes, ~ unclear, or ✗ not on your resume. |
| Improve my chances | Rewritten resume bullets, LinkedIn headline and About section aimed at this job's gaps — using only what you've really done. Invented numbers are flagged; gaps wording can't fix get honest next steps. |
| Job application tracker | Save jobs you applied to, update status (applied → interview → offer), notes, search, filters, CSV export, backups. |
| Were the predictions right? | Compares the predicted chance with your real interview rate, so you can see if the estimates hold up for you. |
| Resume upload | Import your resume from PDF or text. |
| Compare jobs | Every job you check is ranked by your chance, side by side. |
- Sign in with Pollinations — you pay a tiny amount of your own AI credit (Pollen) per check; there's no subscription and no account with us.
- Add your resume once — it's saved only in your browser.
- Paste a job ad and click Should I apply?
Under the hood, one call to Jev — a decision model on Pollinations that returns calibrated probabilities instead of free text — answers typed questions in parallel: pass-the-screen probability, skills overlap, experience, seniority, location fit, deal-breakers (license, degree, clearance, work permit) and one yes/no per requirement. The verdict rules live in code, not in the model. "Improve my chances" uses a separate text model through the OpenAI-compatible /v1/chat/completions endpoint.
Most resume checkers count overlapping keywords (TF-IDF) or ask a generic chatbot for a score. A keyword score tells you which words match; it doesn't tell you your odds. ShouldIApply asks a model built to be honest about probability, shows which requirements you're missing, and refuses to invent experience when it helps you rewrite.
- Your resume, job ads and application history stay in your browser (
localStorage). There is no backend. - Sign-in uses OAuth 2.1 + PKCE; the access token lives only for the browser tab (
sessionStorage). - Only the resume and job ad text are sent to Pollinations when you run a check.
It's a static site — plain HTML, CSS and JavaScript, no build step.
git clone https://github.com/notsointresting/shouldiapply.git
cd shouldiapply
python -m http.server 8000 # then open http://localhost:8000/
npm test # optional: logic checksIt uses a public Pollinations App Key (pk_…) in config.js. To use your own, create one at enter.pollinations.ai/keys and add your site URL (and http://localhost:8000/) as redirect URIs.
- Vanilla JavaScript (ES modules), HTML, CSS — no framework, no build
- Pollinations — Jev decision model (
/alpha/decisions) and text models (/v1/chat/completions), Bring Your Own Pollen sign-in - pdf.js for resume PDF import (loaded only when used)
- Service worker for offline use; hosted on GitHub Pages
| File | Purpose |
|---|---|
index.html |
The page: Check a job · My resume · My applications. |
styles.css |
Styling, light and dark theme. |
config.js |
App Key, endpoints, model names. |
auth.js |
Sign-in (OAuth 2.1 + PKCE). |
jev.js |
The check: one /alpha/decisions call, requirement extraction, verdict rules. |
improve.js |
"Improve my chances": rewrite prompt + invented-number guard. |
tracker.js |
Local application tracker, backups, CSV export. |
app.js |
Controller. |
sw.js, manifest.webmanifest, icon.svg |
Installable + offline. |
test.mjs |
npm test. |
- The chance is an estimate from your resume and the job ad — a guide, not a guarantee. It can't see the other applicants.
- PDF import reads text only; scanned (image) PDFs need to be pasted.
- There's deliberately no auto-apply. Mass auto-applying hurts candidates and gets flagged. ShouldIApply helps you decide.
Ideas adapted (MIT) from hugounoclaw/ats-checker (paste-and-score UX), ParasKoundal/JobTracker (local tracker), and the Profile Optimizer in sergebulaev/linkedin-skills (rewrite rules). Built on Pollinations following their Bring Your Own Pollen guide. Curious about decision models? See awesome-jev-family.
⭐ If ShouldIApply saved you time, star it on GitHub — it helps other job seekers find it. Bugs and ideas: open an issue.