Image to Text — Free Online OCR Tool
Extract text from images in your browser. Free OCR for photos, screenshots, and scans. Nothing uploaded, 11 languages.
Drop an image here or choose a file
PNG, JPG, WebP, BMP, or any image your browser can display
Pull text out of images without uploading anything. Drop a screenshot, photo, or scanned page, pick a language, and the tool runs Tesseract OCR inside your browser to return searchable, editable text. Supports eleven languages including English, Spanish, French, German, Chinese, Japanese, Korean, Arabic, and Russian. Because the OCR engine runs client-side as WebAssembly, private documents and receipts never leave your device.
Optical character recognition turns pixels into characters. That sounds simple, but the hard parts are everywhere: tilted pages, low contrast, background noise, fonts the model has never seen, and languages with thousands of glyphs. This tool uses Tesseract, the open-source OCR engine originally built at HP and now maintained by Google, compiled to WebAssembly so it runs inside the browser tab.
Results are best on clean printed text: book pages, screenshots, PDFs that were exported as images, product labels. The engine reads line by line and returns a block of text plus a confidence score. Scans benefit from being straight and well-lit. If a photo is rotated, cropping and rotating it before running OCR noticeably improves accuracy. Handwriting is a mixed bag. Neat block letters sometimes work, cursive usually does not. Tesseract was trained on typeset text, not handwriting.
Language choice matters more than people expect. Picking the wrong language forces the engine to guess characters from a model that has no idea what the letters look like, and quality collapses. For mixed-language documents, pick the language of the majority of the text. Everything runs locally, so there is no file size limit beyond what your device can hold in memory, and nothing is sent to a server.
- 1
Upload an image
Drag a file onto the drop zone or click to choose one. JPG, PNG, WebP, and most other image formats your browser can display all work.
- 2
Pick the language
Choose the language of the text in the image. The first run in each language downloads a small trained model, then caches it for future use.
- 3
Extract and copy
Click Extract text to run OCR. When finished, review and edit the output, check the confidence score, then copy to clipboard or download as a .txt file.
Scanned documents and PDFs
Pull quotes out of a scanned contract or research paper that was saved as an image instead of searchable text.
Screenshots with text
Grab an error message, chat transcript, or code snippet from a screenshot without retyping it by hand.
Receipts and invoices
Digitize totals, vendor names, and line items from phone photos of paper receipts for expense tracking.
Book pages and study notes
Capture passages from physical books or printed handouts to quote in notes, summaries, or translations.
Is my image uploaded anywhere?
No. The Tesseract OCR engine runs entirely in your browser as WebAssembly. The image and extracted text stay on your device. Nothing is uploaded, logged, or stored on a server.
How accurate is the recognition?
Tesseract works best on clean, high-contrast printed text and typically hits 90 to 99 percent character accuracy in good conditions. Blurry, tilted, or low-light photos drop accuracy. Handwriting is hit or miss: neat block printing sometimes works, cursive usually does not.
Why does the first run take longer?
The first time you pick a language, the browser downloads a trained model for that language (a few megabytes). After that, the model is cached and every subsequent run in the same language starts almost instantly.
Which languages are supported?
English, Spanish, French, German, Portuguese, Italian, Chinese (Simplified), Japanese, Korean, Arabic, and Russian. Each language uses its own model. For best results pick the language of the dominant text in the image.
Can it read handwriting?
Not reliably. Tesseract was trained on typeset text and works best on printed characters. Very tidy block handwriting sometimes works, but flowing cursive, quick notes, and most real-world handwriting produce poor results. For handwriting-heavy documents, specialized HTR tools are a better fit.