Turn photos, scans, and screenshots into editable text you can copy anywhere. Paste a screenshot straight from your clipboard, batch-convert a stack of images, and download it all as a text file — without your images ever leaving your device.
Text trapped in an image is text you can't use — you can't copy a phone number out of a screenshot, search a photographed receipt, or edit a scanned paragraph. This tool runs OCR (optical character recognition) on any image and hands the text back to you, editable and copyable, in seconds.
What makes it different from the usual "image to text" sites is where the work happens: your images are never uploaded. The OCR engine runs inside your browser, which matters because the things people extract text from are often exactly the things they'd least want on someone else's server — receipts, IDs, medical letters, contracts, whiteboards full of company plans, or private chat screenshots.
Ctrl+V (or Cmd+V on a Mac) anywhere on the page and the image in your clipboard is added instantly. Screenshot → paste → copy the text out: done.The tool uses Tesseract, the most widely used open-source OCR engine in the world, compiled to WebAssembly so it runs entirely in your browser. The first run downloads the engine and your chosen language model (a few megabytes, cached for future visits); after that, recognition happens locally — even offline. Small images are automatically upscaled before recognition, which noticeably improves accuracy on screenshots.
Paste the screenshot straight in with Ctrl+V (or drop a photo) and the OCR types out the text, editable and copyable, with a confidence score. The text was always there; now you can select it.
Seventeen, including the major European and Asian languages. Pick the language of the source for the cleanest extraction.
Yes. Batch a set of photos or scans and each produces its own editable text block, with a combined download when you want everything in one file.
Very good on clear print and decent screenshots, and lower-resolution sources are automatically upscaled before recognition to help. Handwriting and stylized fonts remain hard for any OCR, so expect to fix a word here and there.
No. Recognition runs entirely in your browser, which matters because the images people extract text from (documents, IDs, letters) are the ones that least belong on someone's server.