Free tool
Bengali image to text, read on your device.
Khatian and porcha copies, marksheets, notices, newspapers and book pages in বাংলা, read in your browser and given back as text you can paste, search or translate. Numbers and English on the same page come along.
The first time, your browser fetches the reader: about 9 MB, and 1 to 6 MB for each other language. After that, a page takes a second or two, even a tilted phone photo.
Runs in your browser. Your files are never uploaded.
The text appears here, ready to copy or fix.
Several files, or a whole form? Tell Sochen “Read this scanned rent agreement and tell me the notice period and the deposit” and it does every step, checks each against the notification, and still keeps your files on your device.
Try Sochen free →How Bengali reads
- Bengali has many conjuncts (যুক্তাক্ষর) that don't look like their parts: ক্ষ, ঙ্গ, ক্ত, ন্ধ, ষ্ণ. They're single shapes on the page, so a smudged one is read as a different letter; these are the words to check.
- Some vowel signs are written before the consonant (ে, ি) or on both sides of it (ো, ৌ). The reader gives them back in correct Unicode order, so what you copy is spelt right even though the page draws it in another order.
- য় and য, ড় and ড, ঢ় and ঢ differ only by a dot under the letter. A faint dot is dropped and the word becomes a near-miss (like জাতীয় read as জাতীয).
- Bengali digits ০–৯ are common in West Bengal government documents, newspapers and older books. They come through as Bengali digits, not as 0–9.
- Assamese uses the same script with ৰ and ৱ added. For an Assamese page choose Assamese instead; the Bengali model reads those two letters as র and ব.
Bengali documents people bring
- Khatian and porcha (RoR) copies from Banglarbhumi, to read plot numbers, area and the recorded owners.
- Madhyamik and Higher Secondary marksheets and admit cards from the West Bengal boards.
- Notices, circulars and scheme forms from West Bengal and Tripura government offices.
- Rent agreements, affidavits and society letters typed in Bengali.
- Pages of Bengali books, newspapers and study notes.
Getting a better read
- Scan at 300 DPI or photograph the page flat so the letters are at least 20 pixels tall. The dots and the small hooks of conjuncts are what get lost first.
- Choose Bengali on its own. Adding Assamese for a Bengali page makes the reader hesitate on র and ৰ.
- For a Banglarbhumi printout, crop to the table before reading; the ruled lines close to small type confuse the line finder less that way.
- Old Bengali PDFs typed in a legacy font (SutonnyMJ and the like) copy out as gibberish. Choose only Bengali and the reader reads the picture, giving real Unicode.
Worth switching on alongside Bengali: Assamese, only if the page is Assamese. The reader takes up to three languages, and English is read with every Indian language automatically.
Scanned PDFs are read page by page, up to 30 pages at a time; a PDF with its own text layer is copied exactly, without reading the picture.