OCR: recognize text in a PDF or image
Extract the text from a scan or photo (French, English) and create a searchable PDF.
Processed in your browserFree
Related tools
- Extract text from a PDFGet all the text of a PDF as TXT or Markdown, page by page.
- Scan a documentPhotograph your pages with your phone, check them, crop or straighten, then get a multi-page PDF.
- Search inside a PDFFind a word or phrase in a PDF, with the page and context of every match.
- Fix a photographed documentStraighten a sheet photographed at an angle: corner detection, manual adjustment, perspective and contrast correction.
About this tool
Recognize the text in your scanned documents, photos and image-only PDFs with Tesseract (open source), running in your browser: choose the language (French, English or both), process page by page with the preview alongside the text, correct the result, then copy it, download it as TXT or Markdown, or create a searchable PDF that keeps the original image with an invisible text layer.
How to use this tool
- 1Drop a scanned PDF or an image (JPG, PNG, WebP) and choose the language.
- 2Start recognition: the engine is downloaded once, then each page is processed (a few seconds per page).
- 3Proofread and correct the text page by page, then copy it, download it as TXT/Markdown or create a searchable PDF.
Limits to know
- Tesseract 5 engine (LSTM) running in the browser: slower than an online service, but fully private. Expect 2 to 15 seconds per page depending on the device.
- Quality depends on the document: sharp, well-contrasted printed text at 200-300 dpi gives the best results. Handwriting, decorative fonts and blurry photos are poorly recognized.
- Tables and column layouts may get mixed up; stamps and signatures sometimes produce stray characters.
- The searchable PDF keeps the image and adds an invisible text layer: the formatting is not rebuilt. Words with low confidence are ignored.
- Up to 200 pages per document; browser memory limits very large files.
- Accepted formats
- PDF, JPEG, PNG, WebP
- Maximum size
- 100 MB
Privacy and file processing
Photos and recognized text stay in your browser. The OCR engine (Tesseract, WebAssembly) and its language models are downloaded from this site, never your documents.
Read our privacy policyFrequently asked questions
No. Tesseract runs in WebAssembly in your browser. Only the engine and the language model (public files, hosted on AVEROPDF) are downloaded, never your pages.
A PDF whose pages remain images (faithful to the original) but contain an invisible text layer positioned over the words. You can select, copy and search text in it, and it can be indexed by file search engines.