Scalir — Bulk Image Optimiser

Scalir is a free, private, in-browser bulk image optimiser. Point it at images and it will:

• Resize anything larger than 2000px on its longest side down to fit a 2000×2000 box (aspect ratio kept, never cropped).
• Compress anything still over 1 MB to the highest quality that fits, converting to WebP when that yields a meaningfully smaller file.
• Save every result with a scaled_ prefix.

Everything runs on your device — no uploads, no server, no AI, no tracking. Built with Svelte + Vite and the jSquash WebAssembly codecs (MozJPEG, libwebp, libpng) derived from Google's Squoosh.

Live previews

Both long-lived branches have a live deployment:

Branch	Environment	URL
main	Production (stable release)	https://scalir.org
dev	Preview / WIP (latest work in progress, may be unstable)	https://scalir.shad.digital

Use the preview to try features before they ship in a release.

Quick start

npm install
npm run dev      # http://localhost:5173

Build the static site (deployable to any static host):

npm run build      # full landing site → dist/
npm run build:app  # tool only (no marketing pages) → dist-app/
npm run preview    # preview the production build

Run the engine tests:

npm test

Using it

1. Drag images in, Choose files, or (on Chromium browsers) Choose folder….
2. Adjust options if you like — max dimension, max size, prefix, WebP on/off, quality floor.
3. Click Optimise, then Download all (ZIP), download individual files, or (Chromium) Save all to folder….

Browser support

Capability	Chrome / Edge / Opera	Firefox	Safari
Drag-drop / choose files	✅	✅	✅
Download results (ZIP / individual)	✅	✅	✅
Pick a folder / save straight to a folder	✅	—	—

Where the File System Access API isn't available (Firefox/Safari), the app falls back to a ZIP download — no features are lost, just the direct-to-folder convenience.

Docker

Two Dockerfiles are provided — pick the one matching what you want to serve:

File	Builds	Serves	Compose service
Dockerfile	npm run build:app → dist-app/	tool only	scalir (:8080)
Dockerfile.landing	npm run build → dist/	full landing site + tool	scalir-landing (:8081)

# Tool only (most self-hosters):
docker compose up -d scalir          # → http://localhost:8080

# Full landing site:
docker compose up -d scalir-landing  # → http://localhost:8081

Deployment panels (Dokploy / Coolify / CapRover): create an Application (not a Compose service), connect the repo, then set Build Type → Dockerfile — this is a separate card below the Git source section, not inside the GitHub tab. Enter the Docker File path (Dockerfile for the tool, Dockerfile.landing for the full site) and set the container port to 80.

Self-host vs hosted

This repo is the unlimited self-host build. The hosted demo applies a soft per-session image cap via an environment variable:

# .env  (hosted build only)
VITE_FREE_LIMIT=25

Leave it unset for unlimited. Because all processing is client-side, the cap is purely a UX choice — it has no effect on hosting cost, and the self-host build removes it.

Project layout

src/core/        framework-free image engine (resize · compress · naming · EXIF)
  ├─ optimise.ts   the 3 rules, orchestrated
  ├─ rules.ts      format detection, resize math, quality ladder
  ├─ exif.ts       JPEG orientation read + apply
  ├─ codecs.ts     jSquash (WASM) codec bindings for the browser
  └─ naming.ts     scaled_ prefix logic
src/App.svelte   the UI (drag-drop, options, progress, results, downloads)
test/            vitest suite (mock codecs — runs in Node, no browser needed)

The engine is deliberately isolated from the UI and codecs (via a Codecs interface), so it's unit-tested in Node and can later be reused in a desktop (Tauri) build or a CLI.

Roadmap

• Web Worker offload for large batches (keep the UI buttery during big jobs)
• Presets for common use cases (Tiny/WebP, Web, Email-JPEG, …)
• Desktop app (Tauri 2) for Windows & Linux
• AVIF output (@jsquash/avif)
• Drag-to-reorder the queue

Contributing

Contributions are welcome! Please read the Contributing guide before opening a PR, and note our Code of Conduct. For bugs and ideas, use the issue templates. To report a security issue privately, see SECURITY.md.

License

MIT — see LICENSE. Free forever; if it's useful, sponsorship is welcome but never required.