Image to Text
Copy text from photos, screenshots and scanned documents with OCR that runs entirely in your browser. 12 languages, including Arabic.
Your files never leave your device
Typing out text from a photo is slow and easy to get wrong. Optical character recognition (OCR) does the reading for you: it finds the letters in an image and turns them into text you can copy, edit, search or translate. It is useful for receipts and invoices, lecture slides, screenshots of messages, printed forms and pages from a book.
This tool runs the OCR engine inside your browser, so the picture is read on your own device rather than sent to a server. That matters when the image shows a bank statement, an ID card or a private conversation. The first run downloads the engine and the language data, and later runs are faster because your browser keeps them.
For the best result, use a sharp, well-lit image with the text straight and filling most of the frame. Choose the language of the text before you start, including Arabic, and check names and numbers in the result before you rely on them. OCR is very good, but not perfect, on handwriting, low-contrast prints and decorative fonts.
or drop images here, or paste a screenshot (Ctrl+V)
JPG, PNG, WebP, GIF or BMP · up to 50 MB each · several images at once
Text trapped in a picture is useless until you can copy it: a receipt, a slide from a lecture, a screenshot of a message, a page of a book. This free Image to Text converter uses OCR (optical character recognition) to turn that picture into editable text in seconds, and it does it without uploading your image anywhere.
Your files never leave your device. This Image to Text converter runs 100% in your browser. Your images are read on your device and never uploaded; there are no accounts and no tracking inside the tool.
How to extract text from an image
- Add your image. Click Select image, drop one or more images on the box, or paste a screenshot with Ctrl+V (⌘+V on Mac).
- Wait for the first read. The tool reads English straight away. The first time, it downloads the OCR engine and language data, which your browser then keeps in its cache.
- Change the language if needed. Pick the language of the text, such as Arabic or French, and press Read text again. Keep “also contains English” ticked for mixed text.
- Check, copy or save. Recognised words are highlighted on the image. Fix any mistakes in the text box, then Copy text or Download .txt.
Why this OCR tool is different
How OCR works
Optical character recognition finds the shapes of letters in an image and decides which characters they are. This tool uses Tesseract, an OCR engine first developed at Hewlett-Packard, later open-sourced and long maintained with Google’s support. Modern Tesseract uses an LSTM neural network that reads whole lines of text at a time rather than one letter at a time, which makes it much better at real-world fonts and spacing.
Normally that engine runs on a server. Here it runs inside your browser, compiled to WebAssembly. Before recognition, the tool prepares your image: very large photos are scaled down to keep things fast, small images are scaled up so letters have enough pixels, and the optional enhance step converts to greyscale and stretches the contrast so faint text stands out from the background.
Tips for the best results
- Crop tightly. Remove borders, hands and backgrounds so the engine sees only text.
- Shoot straight on. Hold your phone parallel to the page, in good light, without shadows across the text.
- Pick the right language. The language model affects which letters and words are expected; Arabic text read as English will come out as gibberish.
- Use screenshots, not photos of screens. A screenshot of a screen gives near-perfect text; a photo of a monitor adds glare and moiré.
- Check numbers carefully. OCR can confuse similar shapes such as 0 and O, 1 and l, or 5 and S. Low-confidence words are shown in pink.
What you can do with the text
Once your text is in the box you can edit it, copy it into an email or document, or download it as a UTF-8 .txt file that opens on any device. To check the length of what you extracted, paste it into the Word Counter. For PDFs that contain real text, skip OCR entirely and use PDF to Text, which copies the exact characters instead of guessing them from pixels.
Privacy and what is downloaded
When you first run the tool, your browser downloads three things from jsDelivr, a public open-source CDN: the tesseract.js library (checked with a Subresource Integrity hash), the WebAssembly OCR engine, and the language model for the language you chose. The language model is saved in your browser’s local storage so it does not need downloading again. Those downloads contain no information about you or your image. The image itself is only ever read into your browser’s memory, and it is gone when you close or reload the page.
Frequently asked questions
Is this OCR tool free?
Yes. It is free with no sign-up, no daily limit and no watermark, because the recognition runs on your own device.
Are my images uploaded?
No. Text recognition is done in your browser by Tesseract, the open-source OCR engine, compiled to WebAssembly and run through tesseract.js. Only the engine and the language data are downloaded from the public jsDelivr CDN. Your images never leave your device.
Which languages are supported?
English, Arabic, French, Spanish, German, Portuguese, Italian, Turkish, Indonesian, Hindi, Russian and Simplified Chinese. You can combine any of them with English for mixed text.
How accurate is it?
On clear printed text, such as screenshots, documents and book pages, accuracy is usually very high. It drops with blurry photos, unusual fonts, curved pages, very small text and busy backgrounds. Handwriting is generally not recognised reliably. The confidence score and the coloured word boxes help you see where to check.
How can I get better results?
Use a sharp, well-lit photo taken straight on, crop away everything except the text, and choose the correct language. Keep Enhance on for small screenshots and faded scans; turn it off if a photo with coloured text comes out worse.
Can it read text from a scanned PDF?
Convert the PDF pages to images first with our PDF to JPG tool, then drop those images here. If the PDF has real (selectable) text, PDF to Text is faster and exact.
Why is the first run slower?
The OCR engine and the language model are a few megabytes each. They download once: the engine stays in your browser’s cache and the language model is saved in your browser’s own storage on this device, so later runs start much faster.
Does it keep the layout or formatting?
No. The output is plain text with line breaks, which makes it easy to paste anywhere. Tables come out as lines of text, and columns are read block by block.
Related tools
By Abdullah Al-Janoubi · WorldScope Tools · Updated October 2026
Comments