Free OCR tool
Drop in a photo of a document, whiteboard, receipt, or sign — or a plain screenshot — and copy the text out of it. OCR runs on your device: no upload, no signup, no limits.
Drop a picture here
or
PNG, JPG, WebP · English text works best
Runs in your browser with open-source OCR (Tesseract via WebAssembly). Your image never leaves your device.
Extracted text
Links found
Need the link behind a title with no visible URL? That is what SnapAction does.
This tool runs the same on-device OCR as SnapAction's screenshot to text converter, tuned for a broader kind of input: a picture from your camera roll, not only an app screenshot. Point it at a photographed page, a whiteboard from a meeting, a business card, or a printed receipt, and it reads the pixels back into copyable text inside your browser tab — no server, no account, no upload.
Photos behave differently from screenshots. A screenshot is already flat, sharp, and evenly lit; a photo can be angled, shadowed, or catching glare. For the cleanest result, shoot straight-on rather than at an angle, fill the frame with the text, and avoid glare on glossy paper or screens. If a picture comes back empty or garbled, retaking it closer and straighter usually fixes it faster than running the same crooked shot through OCR again.
Copying text out of a picture gets you the words on the page. If what you actually photographed was a saved iPhone screenshot — a repo, a listing, a booking, an event flyer — SnapAction's screenshot AI can go a step further: it identifies the resource the screenshot represents, recovers the canonical link with AI analysis and web search when one isn't visible, and saves the result as a typed card with a next action. That part is built specifically for the iPhone screenshot library via PhotoKit, not for general camera photos.
See the mechanics in OCR vs resource extraction, or start with the extract links from screenshots workflow.
Drop the picture into the tool above or choose a file. OCR runs in your browser and returns the text, ready to copy or download — no upload required.
Yes. The OCR engine reads any picture — a photographed document, whiteboard, sign, business card, or receipt — the same way it reads an app screenshot. It reads pixels, not the source of the image.
Photos taken at an angle, in low light, or with glare are harder for OCR than a flat screenshot. Retake the photo straight-on, fill the frame with the text, and avoid reflections for a cleaner result.
No. Recognition happens locally in your browser via WebAssembly (Tesseract). Your picture never leaves your device.
Yes — screenshots, photos, scans, and most other image formats all go through the same on-device OCR.