Bijoy Converter
বাংলা
Online

Image / PDF to Text (Bangla OCR)

Upload a photo, screenshot, scan or PDF and get editable Bengali and English text back. Recognition runs in your browser with Tesseract OCR, so your documents stay private.

Drop an image or PDF here

JPG, PNG, WebP up to 10.00 MB · PDF up to 20.00 MB and 30 pages

OCR runs on this device with Tesseract. Your image or PDF is never uploaded. The language model is downloaded once and cached.

How to extract text from an image or PDF

  1. 1Drop a photo, screenshot, scan or PDF onto the box. Straight, sharp, high-contrast images work best.
  2. 2Choose the language: Bengali + English for mixed documents, or a single language for slightly better accuracy. Click Extract text.
  3. 3Review the text, fix any mistakes in the box, then copy it or download it as TXT or PDF. Use the Text Fixer for broken vowel signs.

About this Bangla OCR tool

Optical character recognition turns the pixels of a scanned page or photo into real, searchable Unicode text. This tool runs Tesseract OCR with its Bengali and English models inside your browser, so NID cards, certificates, bank papers and old newspapers are processed on your own device rather than on a server.

Bengali OCR is harder than English because of conjuncts and vowel signs, so expect a few mistakes on low-quality scans. Printed text in common fonts at 300 DPI recognises well; handwriting is not supported. The output is Unicode, so you can convert it to Bijoy with our converter if you need SutonnyMJ.

Features

  • ✓Bengali, English or both at once
  • ✓JPG, PNG, WebP images and multi-page PDFs
  • ✓Editable result with confidence score
  • ✓Copy, download as TXT or Bengali-safe PDF
  • ✓In-browser Tesseract OCR, no upload
  • ✓Model cached after the first run, works offline afterwards

Frequently asked questions

How accurate is the Bengali recognition?

Clean printed text at 300 DPI typically comes out 90% or better. Blurry phone photos, skewed scans, decorative fonts and Bijoy text rendered in a Unicode font reduce accuracy. Cropping to the text area helps.

Can it read handwriting or Bijoy fonts?

Handwriting is not supported. Bijoy fonts such as SutonnyMJ are printed Bengali glyphs, so the OCR reads them as Bengali and outputs Unicode; the result may need a pass through the Text Fixer.

Why does the first run take longer?

The browser downloads the Bengali and English models (about 2.5 MB) and the OCR engine on the first run. They are cached, so later runs start immediately and even work offline.

Related Tools

Related guides