How to extract text from an image
- Drop, browse or paste an image containing text — a screenshot, scanned page or photo.
- Wait a moment while the OCR engine reads it (the first run downloads the engine once, then it's cached).
- Review and edit the recognized text, then click Copy text.
OCR that respects your privacy
Optical character recognition turns pixels of text into characters you can copy, search and edit. Most online OCR tools upload your image to their servers to do it — a problem when the image is a receipt, an ID, a contract or an internal screenshot. This tool runs the Tesseract engine directly in your browser through WebAssembly, so the picture stays on your device and only the engine itself is fetched (once) from a CDN. It works best on clear, high-contrast printed or typed text: screenshots, scanned documents, book pages, signs and receipts. Handwriting and decorative fonts are hit-or-miss, and a sharper image always beats a blurry one.
Frequently asked questions
How does image-to-text work without uploading?
The Tesseract OCR engine runs inside your browser via WebAssembly. Your image is read locally and recognized on your device — it's never sent to a server. The engine downloads once from a CDN.
What kinds of images work best?
Clear, high-contrast printed or typed text — screenshots, scans, book pages, signs, receipts. Handwriting and stylized fonts are unreliable.
Which languages are supported?
English. Tesseract supports many languages, but each adds a separate download, so this page focuses on English for speed.
Why is the first run slower?
The first image triggers a one-time ~4 MB download of the engine and language model. After that it's cached and much faster.