Image to Text
Read the words out of a screenshot, scan or photo of a document.
- Add
- Adjust
- Make
- Keep
How to use this tool
- Add your files — drag them onto the page or choose them. They stay on your device.
- Say what is in the image — a full page, one block, a single line, or scattered words — so the recogniser scans the right way.
- Leave Clean the image up first on for photos and faint scans; switch Save as to Markdown when paragraphs matter.
- Press “Read the text” and download the results, individually or as one ZIP.
About this tool
Read the words out of a screenshot, scan or photo of a document. Free and unlimited, and nothing is uploaded: it all runs in your own browser. Recognition runs entirely on your device — nothing is uploaded to any server, and the result is generated locally before downloading straight from your browser.
Privacy & limitations
Recognition is on-device Tesseract: strong on printed text and clean screenshots, weaker on handwriting, heavy stylisation and low-light photos. The first run downloads the English recognition data (a few MB, once, from this site) — it is cached for offline use afterwards.
Last updated:
Frequently asked questions
How accurate is it?
On a clean screenshot or a straight scan, very. On a photograph taken at an angle in bad light, much less so. The confidence figure it reports is a fair guide to whether the result is worth trusting.
How do I get a better result?
Give it a larger, straighter, higher-contrast image. Cropping to just the text helps more than anything else, because it stops the recogniser trying to read the surroundings.
Does it handle handwriting?
No. The engine is trained on printed text and will produce nonsense from handwriting.
Is my document uploaded?
No. The recognition engine and its language data are downloaded to your browser and run in the page, which is the whole point for anything with personal information on it.