Drop one or more images, or paste a screenshot with Ctrl+V. Choose the language of the text, then press Extract text. The OCR engine reads each image in your browser and shows the text next to a preview. Edit it, copy it, or download everything as a .txt file. Nothing is uploaded.
- Are my images uploaded?
- No. The OCR engine runs inside your browser as WebAssembly, so the image and the text stay on your device. English works fully from this site. For other languages, your browser downloads that language's data file once from a public CDN; only the data file is downloaded, never your image.
- How accurate is it?
- Best on clean, straight, high-contrast printed text such as screenshots, scanned documents and book pages. Handwriting, curved or tilted text, small low-resolution photos and busy backgrounds give weaker results. Check the text and fix mistakes in the box before you copy it.
- Why is the first run slow?
- The first run downloads and starts the OCR engine and language data, a few megabytes. Your browser caches them, so later runs start faster. Reading a page then takes a few seconds, depending on your device.
- Which languages are supported?
- English is built in. You can also pick Hindi, Telugu, Tamil, Bengali, Spanish, French, German, Portuguese, Italian, Russian, Arabic, Chinese, Japanese and Korean. One language is used per run.
- Can I read a PDF?
- Not yet. Take a screenshot of the page or export it as an image, then drop that image here.
