Image tools
Image to text (OCR)
Photo, scan or screenshot in — editable text out. Runs on your device.
Drop an image here
JPG, PNG or WebP · nothing is uploaded
Tip: you can also paste a screenshot with Ctrl+V or ⌘V.
The language file is fetched once from a public CDN (a few MB). The image itself never leaves your device.
How it works
- 1Choose a photo or screenshot of the text, drag it onto the box, or paste it from the clipboard.
- 2Pick the language the text is written in, then press Extract text. Recognition runs inside your browser.
- 3Fix anything it misread in the text box, then copy it or download it as a .txt file.
Specification
- Input
- JPG, PNG, WebP, BMP — photos, scans or screenshots
- Languages
- English plus 20 others, selectable
- Output
- Editable text — copy it or save as .txt
- Engine
- Tesseract, running in your browser
- Uploads
- None — only the public language file is fetched
Questions
How accurate is it?
On a clear, straight, well-lit photo of printed text it is very good — typically only the odd character in a page. Accuracy drops with blur, skew, low contrast, tiny type and decorative fonts. Crop to the text, keep the lines horizontal and use the largest, sharpest version you have.
Can it read handwriting?
Not reliably. The engine is trained on printed text; neat block capitals sometimes work, cursive almost never does.
Is my image uploaded anywhere?
No. Recognition runs inside your browser with Tesseract compiled to WebAssembly. The only download is the public language file for the language you pick — a few megabytes, fetched once from a CDN and then cached. The image itself never leaves your device.
Why is the first run slower?
The engine and the language data have to be fetched the first time — usually 5–15 MB depending on the language. After that they are cached and later images are read in a couple of seconds.
Can I extract text from a PDF?
Convert the page to an image first with the PDF to JPG tool, then read that image here. A PDF that already contains real text (one you can select) does not need OCR at all — copy the text straight from your PDF viewer.
Which languages are supported?
English, Spanish, Portuguese, French, German, Italian, Dutch, Indonesian, Turkish, Polish, Russian, Arabic, Hindi, Urdu, Chinese (simplified and traditional), Japanese, Korean, Vietnamese and Thai. Pick the language the text is written in — the wrong language file gives noticeably worse results.
Is this the same as "photo to text"?
Yes. It also gets called "image to text converter" and "ocr online free" and "extract text from image". They all describe this one job, and this page does it.