Image to Text (OCR)
Pull the words out of a screenshot without sending it anywhere.
The recognition runs in this browser. The language data is served from this site and cached after the first run; the picture itself is never sent anywhere.
The first run downloads about 6 MB of recognition data, then caches it.
If the lines are broken up oddly: Line Break Remover
Printed text reads well. Handwriting, low light and heavy blur do not.
The picture stays on this device.
How to use
- Choose a picture or paste one, and pick the languages it contains.
- Press Start. The first run downloads the recognition data, which is then cached.
- Copy the text or save it as a .txt file.
Examples
A screenshot of an error message
Reading the text out makes it searchable, which a screenshot never is.
A photo of a printed page
Printed text reads well. Hold the camera square to the page and the lines come out straighter.
A photo from an iPhone
HEIC photos have to be converted first; the converter here does that in the browser too. HEIC to JPG
Text that comes back in fragments
Recognised text keeps the line breaks of the image. Joining them makes it read as sentences again. Line Break Remover
FAQ
Is my image uploaded?
No. Only the language data is downloaded; your image never leaves your device.
Why is the first run slow?
About 6 MB of recognition data downloads once, then it is cached.
Does it read handwriting?
Printed text works well; handwriting and blurry photos are less accurate.