OCR2textRuns on your deviceCamera to text
The rear camera on a phone, the built-in webcam on a laptop, or an external camera — the same workspace, with document framing, torch control and multi-page capture.
Permission is only requested when you choose the camera, never before.
Guides show the capture area. Turn the torch on where the device supports it.
Capture, then keep going — a letter, a form, a chapter.
Drag the corners onto the page and it is warped back into a rectangle.
All pages at once, combined in the order you captured them.
Plenty of OCR sites tell desktop visitors to go and find their phone. OCR2text opens your webcam instead, with a large preview and a device picker when more than one camera is connected. It is genuinely useful for a page held up in front of the screen, a book, or a label on a box.
A page photographed at an angle produces text lines that fan across the frame, and recognisers cope badly with that. Flatten page lets you put four handles on the page corners; OCR2text solves the homography and warps it back to a rectangle, as though it had been scanned flat.
Where the page sits on a darker surface, the corners are detected for you. When the detection is not convincing, OCR2text says so and hands you the handles rather than quietly applying a bad guess.
A denied permission is not a dead end here. You are offered upload and paste immediately, and everything after that point behaves identically. Camera access can be re-granted from your browser’s address bar at any time.
No account, no upload, no waiting for a server. Open the workspace and drop a page in.