OCR · read text off an image

Drop images here

choose images or a whole folder · paste · snap a photo

Use AI reader (most accurate) — sends the cropped region to a vision model (Gemini); needs configuring

Expected format — clean & group known keys (optional)

Enable Groups × size Separator Fix look-alikes via alphabet Alphabet

Remove characters your keys never use (e.g. delete O so an OCR’d “O” becomes 0). Expected length = groups × size.

Runs entirely in your browser · OCR by ocrs (Rust → WASM) · images never leave your device

Results