Image to Text (OCR)
Pull the words out of a picture
Image
Result
Nothing recognised yet
Drop in an image and press recognise — the text appears here.
Handwriting and heavily angled text are recognised poorly.
Text locked inside an image cannot be searched, copied or corrected. OCR reads the shapes and turns them back into characters. Screenshots and flat scans come out close to perfect; angled photos and handwriting are much harder.
How to use
- 1 Drop in an image containing text.
- 2 Choose the language — picking the right one makes a large difference.
- 3 Press recognise and wait.
- 4 Correct anything that came out wrong, then copy.
Frequently asked questions
Why is the first run so slow?
The recognition engine and language data have to be downloaded first, which takes tens of seconds. After that they are cached and subsequent runs take a second or two.
How do I get better accuracy?
Use the highest resolution you have, keep the text horizontal, and crop away everything that is not text. Straight-on scans and screenshots are near perfect; photos taken at an angle in poor light are unreliable.
Does it read handwriting?
Barely. The engine is trained on printed type. Neat block capitals sometimes work; ordinary cursive does not.