How to use Image to text
- Choose one or more images: photos, screenshots or scans.
- Tick the languages in them.
- Select Extract text.
- Copy the text, or download it as a .txt file.
Copy text out of any picture
Screenshots of error messages, photos of a whiteboard, a picture of a receipt, a scanned page, a slide from a recorded meeting: the text is there, but you can't select it. This tool reads it with OCR, optical character recognition, and gives it back to you as plain text to copy, edit or save.
It uses Tesseract, a widely used open-source OCR engine, compiled to run inside your browser. Add several images at once and the text from each comes back in order, labelled with its file name.
Nothing is uploaded
Most image-to-text sites send your picture to a server. Here the recognition happens on your device: the engine and the language data download once from this site, about 4 to 7 MB depending on the languages, and your browser keeps them for next time. The images never leave your device and nothing is stored.
Languages
English, Hindi, Spanish, French, German and Portuguese are available. Pick the ones in your image, up to three at once for mixed text such as a bilingual form. Fewer languages read faster and more accurately.
Tips for better accuracy
Use the sharpest image you have, taken straight on in good light; crop away anything that isn't text if you can. Screenshots read almost perfectly. Printed text in photos reads well; handwriting, very small text, curved pages and stylised fonts don't. When the engine isn't confident about an image, the tool says so, so you know to check that one. Photos are turned upright using the orientation your camera recorded before they're read.
For a scanned PDF rather than separate images, use OCR PDF, which also makes the PDF itself searchable.
Common uses
Pasting the text of an error dialog into a search or a support ticket, turning a photographed recipe or notice into something you can edit, copying an address or reference number from a screenshot, pulling quotes out of slides, or getting the words from a scanned letter into an email reply without retyping them.
Questions
- Which languages can it read?
- English, Hindi, Spanish, French, German and Portuguese, up to three at once for documents that mix them. Each language's data downloads once, from this site, the first time you use it.
- Are my images uploaded?
- No. The text is recognised in your browser by Tesseract, an open-source OCR engine. Nothing is sent to a server and nothing is stored.
- Can it read handwriting?
- Not reliably. Tesseract is trained on printed and typed text. Neat block capitals sometimes work; joined-up handwriting usually doesn't.