AkuruLiyoby devTuskersOpen editor →

Sinhala OCR — Image to Sinhala Text

සිංහල OCR — රූපයේ අකුරු වචන බවට හරවන්න.

Drop a Sinhala photo, scan, or screenshot. The text appears below in editable Sinhala Unicode. The OCR runs entirely in your browser — your image never leaves your device.

Try AkuruLiyo →

Drop a Sinhala image here, or click to pick a file.

JPEG, PNG, or WebP. Browser-local — your image never uploads.

What this can extract

  • Printed Sinhala from books, magazines, and newspapers
  • Sinhala signage and screenshots
  • Scanned PDFs (one page at a time as an image)
  • Mixed Sinhala + English documents

Handwritten Sinhala is hit-and-miss — Tesseract is trained on printed glyphs. For best results, use the highest-resolution version of the source you have.

How it works

AkuruLiyo runs Tesseract.js in your browser with the Sinhala-trained language model (sin.traineddata, maintained by the Tesseract open-source community). On the first OCR you'll see a one-time ~10 MB download for the language model — subsequent runs are instant because the model lives in your browser's IndexedDB cache.

Nothing is uploaded. We don't have a backend for this — by design, the OCR pipeline never sees a server. The image you drop into the page stays in your tab; close the tab and it's gone.

Frequently asked questions

Is this Sinhala OCR free?
Yes — completely free, no daily limits, no signup. It runs in your browser; we don't even see what you OCR.
How accurate is it?
Excellent on clean printed Sinhala (books, articles, signs); good on photographed pages with reasonable lighting; modest on handwriting or very low-resolution sources. The cleaner the source, the better the output. Mixed Sinhala + English pages work — Tesseract handles both scripts.
Why does the first run take a while?
The first OCR downloads the Sinhala language model (sin.traineddata, ≈10 MB) into your browser's IndexedDB. After that, the model is cached and every subsequent OCR starts in milliseconds. The download only happens once per browser.
Can I edit and format the extracted text?
Yes. Click Open in editor → after OCR and the text drops into a fresh AkuruLiyo document where you can fix any glyph errors, add headings, lists, formatting, images, and export to PDF / Word / Markdown.
Can I OCR a PDF?
v1 accepts image files (JPEG, PNG, WebP). For a PDF, export or screenshot each page as an image and OCR them one by one. Bulk PDF OCR is a candidate for v2 once we add a backend surface.