Document Scanner
Photo → cropped scan + searchable PDF & text. On-device OCR.
Runs entirely in your browser — nothing you enter is uploaded, logged, or tracked.
A real document scanner that runs 100% on your device — take a photo of a page, receipt, contract or whiteboard and it finds the document, straightens the perspective into a flat top-down scan, and cleans it up (colour, greyscale, or crisp black-and-white). Then an on-device OCR engine reads the text so you can export a searchable PDF (the text is selectable and findable, layered invisibly over the scan), a clean image, or the plain text to copy. Drag the corners to fit any angle. Nothing is uploaded — the photo, the scan and the recognised text never leave your browser, and the export carries no GPS or camera metadata.
Export
A PDF that looks like the scan but with selectable, searchable text on top.
Rebuilt on your device — no GPS or camera metadata, no watermark.
Frequently asked questions
Is my document uploaded anywhere?
No. The edge-detection, perspective correction and OCR all run inside your browser on your own device — there are no servers, no upload, no analytics. After the one-time engine download you can disconnect from the internet and it still works. For private documents (IDs, contracts, receipts) that matters: cloud scanners send your pages to their servers; this one never does.
How does OCR run in my browser with nothing uploaded?
It uses Tesseract — a mature open-source OCR engine — compiled to run on your device via WebAssembly. The first time you scan, the engine and English language data download once (a few MB), cache, and from then on every scan is read locally and offline.
What is a “searchable PDF”?
It’s a PDF that looks like your scanned image but has the recognised text layered invisibly on top, aligned to where each word appears. So you can select, copy and search the text (Ctrl/Cmd-F) even though you’re looking at a photo of the page — exactly what professional scanners produce.
How do I get a straight scan from a photo taken at an angle?
The tool auto-detects the document’s four corners; you drag any corner to fine-tune the fit, then it warps that quadrilateral into a flat, top-down rectangle — removing the angle and keystone distortion. Pick black-and-white for crisp text or colour to keep the original look.
How accurate is the text recognition?
It’s good on clear, well-lit printed text in English — the black-and-white “enhance” mode and a straight crop help a lot. Handwriting, very small or low-contrast text, unusual fonts and heavy skew are harder for any OCR. You can always edit the recognised text before exporting.
Does it remove EXIF / GPS metadata?
Always. The scan image and PDF are rebuilt from raw pixels, so the original photo’s GPS location, camera model and EXIF tags can’t carry over.