Private — processed on your device, never uploaded
Image to text
Read the words out of a picture or screenshot.
Language
Tick more than one for mixed text. Each is fetched once and then cached by the browser — 15 MB for what is selected, and nothing after that.
About this tool
A photograph of a page has no text in it, only pixels that look like letters. This reads them: Tesseract, the open-source recogniser, compiled to WebAssembly and run right here. Point it at a screenshot, a receipt, a whiteboard or a scanned page and you get the words back, ready to copy or save. The language data downloads once and is then cached — after that it works offline, and your picture never leaves the device either way.
Private by design — your files never leave your device. Toolpit has no upload step and no server that can see your document — the work happens in this tab, on your machine, and stops the moment you close it.
Related tools
Every one of these runs in your browser too — no upload, no signup, no watermark.