How do I extract text from an image (OCR)?
Image OCR reads the text inside a photo, screenshot, or scan and gives you back a plain string you can copy, search, or paste anywhere.
Drop a JPG, PNG, or WebP (up to 10 MB), pick a language, get the recognised text with a confidence score per block.
Recognition runs on our server using Tesseract - the same engine Google built for Android - with English, Polish, German, French, and Spanish language packs.
Best for clean scans, screenshots of dialog boxes, invoices, receipts, and printed pages. Handwriting and busy backgrounds will struggle.
How to use it
- Drag your image onto the dropzone or click "Choose file" - JPG, PNG, WebP are accepted, HEIC is not (convert it first with the HEIC converter).
- Pick the language that matches your image. Mixing languages on one page works poorly - use the dominant one.
- Click "Extract text". The first run downloads a ~10 MB language pack on the server, so the very first call may take 5 to 15 seconds; later calls are faster.
- Read the extracted text in the box on the right. Use "Copy" to put it on the clipboard or "Download" to save a `.txt` file.
- Toggle "Show word boxes" to overlay every recognised word on the image - useful for spotting missed regions or low-confidence patches.
When this is useful
Where OCR pays off - typical situations:
- Quoting a screenshot in a doc or chat without retyping it word by word.
- Pulling a phone number, email, or address off a photo of a business card or a printed flyer.
- Reading a receipt to track an expense - the totals and line items become searchable text.
- Lifting text from a UI when a developer or designer ships you a flat PNG with no editable layer.
- Old invoices and contracts that were scanned to PDF and lost their text layer along the way.
- Memes, signs, posters - quickly grab the slogan or caption.
If your file is a PDF that already contains a text layer (most PDFs from Word / Pages / Chrome "Save as PDF" do), use the PDF text extractor instead - it is instant, perfectly accurate, and free of OCR errors. OCR is only the right tool when there is no real text in the file, only pixels.