Drop your image here
or click to browse — PNG, JPEG, WebP supported
Read text from photos, screenshots, and scanned documents. Free, private, runs in your browser.
Drop your image here
or click to browse — PNG, JPEG, WebP supported
Unlike other background removers, Clear Canvas processes everything locally in your browser — your photos are never uploaded anywhere.
AI model runs in-browser via ONNX Runtime and WebAssembly — not on a remote server.
Files stay in browser memory and are discarded when the tab closes.
No account needed. Your images are never stored, shared, or sent to any server.
After initial model load, no network connection needed.
Need programmatic access? Use the x402 API for automated processing.
curl -X POST "https://x402.clearcanvas.app/v1/image/ocr?lang=eng" \
-H "Content-Type: image/png" \
--data-binary @screenshot.png| Parameter | Type | Required | Description |
|---|---|---|---|
| lang | string | No (eng) | Language: eng, chi_tra, chi_sim, jpn, kor, spa, deu, fra, por |
POST /v1/image/ocr — Paid with USDC on Base via x402. No accounts needed.
English, Chinese (Traditional & Simplified), Japanese, Korean, Spanish, German, French, and Portuguese.
Accuracy depends on image quality. Clean text on a white background typically achieves 90%+ confidence. Handwriting and complex layouts may have lower accuracy.
No. OCR runs entirely in your browser using tesseract.js (WebAssembly). Only the language model file is downloaded on first use.
These side-by-side pages are the fastest way to see whether Clear Canvas fits your workflow.