Image to Text
Pull the text out of any screenshot, scan, or photo.
Screenshots, scans, receipts, and photos. English OCR.
The first run downloads a ~3 MB English model once, then it is cached by your browser.
Tesseract.js runs in a Web Worker. Your image is never uploaded.
What this tool does
Screenshots are where useful text goes to become unselectable. This tool runs optical character recognition in a background worker in your browser and gives you the words back as editable text.
It is built for the ordinary cases: a screenshot of a receipt, a slide photographed at a conference, a scanned letter, a paragraph trapped in an image. Your image is not uploaded to a server, and there is no account to create.
Nothing uploaded
Tesseract.js runs locally in a Web Worker. Your images stay on your device.
See what it sees
Image preview, word confidence, and a clean text result side by side.
Useful output
Copy the text or download a .txt file. Line breaks are kept where OCR finds them.
How to use Image to Text
- 1
Add an image
Drop a PNG, JPG, or WebP screenshot, scan, or photo. Up to 10 MB.
- 2
Run OCR
The first run downloads the English language model once, then caches it.
- 3
Fix and use
Edit any misread words in the result box, then copy or download the text.
Example
A phone screenshot of an itemized receipt with a merchant name, date, four line items, tax, and total.
Blue Bottle Coffee Nov 4, 2026 Latte 5.50 Croissant 4.25 Subtotal 9.75 Tax 0.85 Total 10.60
FAQ
Image to Text questions
How accurate is the OCR?
For clean screenshots and straight-on scans, accuracy is usually very high. Photos with glare, small text, handwriting, or unusual fonts are harder. You can edit the result before copying it.
Which languages are supported?
The hosted version ships with English. Tesseract supports more than 100 languages; a self-hosted copy can add them by dropping the matching traineddata file into the language folder.
Can it read handwriting?
Occasionally, but it is not reliable. Tesseract is designed for printed text. For handwritten notes, a vision-language model will do better.
Is my image uploaded?
No. The OCR model runs in your browser. The only network request is the one-time download of the language model from this site.
What image formats work?
PNG, JPG, WebP, BMP, and GIF. Very large images are downscaled in the browser before recognition to keep things fast.