ଓଡ଼ିଆ OCR

Odia OCR: turn a photo or scan into editable text

Upload a picture or PDF of printed Odia and get text you can copy, edit and search. Everything runs in your browser, so your file is never uploaded to our server.

Image or PDF → Odia text Free · no sign-up · private
1. Choose a file
or drop a file here You can also paste a screenshot with Ctrl+V JPG, PNG, WebP or PDF · up to 15 MB
0 characters
How to use it

Three steps

  1. Choose your file

    Pick a photo, a screenshot or a scanned PDF. The first time, the Odia reading data (about 1 MB) is downloaded, which can take a few seconds.

  2. Press Read the text

    The page is read on your own device. A multi-page PDF is read page by page, up to 20 pages.

  3. Check and copy

    Correct any mistakes in the box, then copy the text or download it as a Word (.docx), PDF or .txt file.

Better results

Tips for a clean reading

  • Use a straight, well-lit photo with no shadows. Hold the phone parallel to the page.
  • Crop to the text before uploading. Borders, stamps and photos confuse the reader.
  • Printed text in a normal size works best. Scans of at least 200 dpi are ideal.
  • Use Single column or poem for most pages. Choose Page with columns or tables only for newspapers, forms and tables.
  • Turn on Enhance contrast for faded photos. Leave it off for crisp screenshots, and do not enlarge the picture first, because that makes reading worse.
  • Choose Odia + English if the page mixes both languages.

Please check the result. OCR is never perfect. In our tests plain printed Odia was read almost perfectly, while bold or rounded fonts had roughly 1 wrong letter in 15. Handwriting, decorative fonts, very small text and old newspapers are read poorly, and similar-looking Odia letters can be confused. Read the output once before you use it for anything official.

FAQ

OCR questions

Is my file uploaded to your server?
No. The reading happens inside your own browser. Your file never leaves your device. Only the reading engine and the Odia language data are downloaded from a public content network.
Does it read handwriting?
No. It is built for printed text. Handwritten Odia will give poor results.
Why does it take long the first time?
The reading engine and the Odia language data are downloaded once and then kept by your browser, so later readings start faster. Reading itself can take several seconds per page on a phone.
The text came out as strange symbols. What is wrong?
The file may use an old Odia font such as Akruti, which was typed in a different way. If you already have such text, use our Font to Unicode converter. For a scan, retake the picture more clearly and try again.
Can I read a PDF with many pages?
Yes. Up to 20 pages are read in one go. For longer files, split the PDF and read it in parts.

Credits. Text recognition uses Google Cloud Vision when it is available, otherwise Tesseract.js (Apache 2.0), based on the Tesseract OCR engine and its Odia language data, and PDF pages are drawn with PDF.js (Apache 2.0). Both are loaded from a public CDN when you use this page.

Need to fix or convert the text?

Retype it with the Odia typing tool or convert old-font text to Unicode.

Odia typing tool → Font to Unicode
Scroll to Top