Skip to content

About

Private, browser-based batch image upscaling and fitted-curve SVG vectorization.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Latest commit

 

History

1 Commit

Folders and files

Repository files navigation

PixelForge Batch

PixelForge Batch is a dependency-free browser app for processing several images in one session.

Run it

Serve this folder over HTTPS (or localhost during development) and open it in a current Chrome or Edge browser. Static hosting is sufficient; no application server is required. The included manifest and service worker provide an installable/offline application shell.

What it does

  • Queues up to 30 PNG, JPG, WebP, BMP, GIF, or SVG files.
  • Accepts multi-file and folder selections, locks the queue until every file is hashed and decoded, and skips only exact-content duplicates.
  • Upscales raster output by 2×, 4×, 6×, or 8×.
  • Offers multi-pass smoothing, single-pass smoothing, and nearest-neighbor pixel-art scaling.
  • Adds optional edge sharpening for outputs within a safe memory threshold.
  • Exports PNG, JPG, or WebP files.
  • Creates fitted-curve SVG traces with configurable colors, detail, and smooth/corner geometry.
  • Preserves and sanitises existing SVG geometry instead of rasterising and retracing it.
  • Runs heavy raster and vector work in a cancellable Worker with OffscreenCanvas.
  • Adds 72, 96, or 300 DPI metadata to PNG/JPG output and supports a custom JPG background.
  • Guarantees collision-safe output names, offers before/after previews, and can save to a selected folder where supported.
  • Preflights queued dimensions, disables unsafe scale choices, and automatically selects a safe lower scale when available.
  • Uses an 8192 px desktop edge ceiling (4096 px on iOS-like devices), updates the selected scale before processing, and keeps dimensions, labels, and filenames consistent with the effective scale.
  • Preserves a successful raster or vector file when the other requested output cannot be created.
  • Saves processing settings in local browser storage.
  • Processes everything locally; images are not uploaded.
  • Detects file:// launches and explains that localhost or HTTPS is required for the vector Worker.

Best results

  • Use Multi-pass smooth for photos and general artwork.
  • Use Pixel art / nearest for sprites and intentionally hard pixel edges.
  • SVG tracing works best for logos, icons, line art, and flat illustrations.
  • Start with Balanced vector detail. Higher detail can produce much larger SVG files.

Honest limitations

  • This is high-quality browser resampling, not generative AI super-resolution. It smooths scaling and sharpens edges but cannot recreate detail absent from the source.
  • SVG photo traces are stylized, palette-reduced curve approximations, not photorealistic reconstructions.
  • Animated GIFs are processed as a single decoded frame.
  • Raster output is capped at 8192 × 8192 on the supported desktop path and 4096 × 4096 on iOS-like devices. Canvas limits vary by browser and hardware, so PixelForge also keeps its preflight and encoding failure checks.
  • Edge sharpening is skipped above 12 million output pixels to limit memory use.
  • “Download all” starts separate browser downloads; the browser may ask for multi-download permission. Compatible Chromium browsers also offer direct output-folder saving.
  • Canvas export cannot preserve the source ICC profile, EXIF, XMP, bit depth, animation, or RAW/HEIC/TIFF data. DPI metadata is newly written for PNG/JPG, but full metadata/color-profile preservation needs a future WASM codec.
  • This build performs faithful high-quality resampling, not AI super-resolution. A model-backed WebGPU/WASM engine remains a separate future upgrade.

About

Private, browser-based batch image upscaling and fitted-curve SVG vectorization.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages