BRev Image to text (OCR) — free
Image to text (OCR) for screenshots, scans and photos — English, Arabic and more, read in your browser and never uploaded.
Drop an image here
Screenshot, scan or photo · text is extracted on your device
Choose imageThis image-to-text tool reads the words in a screenshot, scanned page or photo with Tesseract, the open-source OCR engine, compiled to WebAssembly and running in your browser tab. The image you choose is read from your own disk and is not sent anywhere: the engine and its language data are the only things the page downloads, and both come from this site rather than a third-party CDN. Five languages are available, English, Arabic, French, Spanish and Hindi, and you can switch on several at once for an image that mixes them, such as a receipt printed in English and Arabic. Each language's recognition data is fetched the first time you use it, from 1.4 MB for Hindi to 10.9 MB for English, and is then kept in your browser's storage. Reading starts as soon as you choose an image, in English unless you have picked other languages. The result lands in an editable box with character and line counts, ready to copy or save as a .txt file.
What you can do
- Tesseract OCR (tesseract.js 7) running in a WebAssembly worker, using its LSTM neural-network recogniser
- Five languages: English, Arabic, French, Spanish and Hindi, in any combination for one read
- Reading starts as soon as you choose or drop an image, with a progress bar for each stage
- Engine and language data are served from BRev's own site and kept in the browser after first use: English 10.9 MB, Spanish 8.4 MB, French 6.3 MB, Arabic 1.7 MB, Hindi 1.4 MB
- With Arabic selected, the result box uses automatic text direction so right-to-left lines display correctly
- Editable result with character and line counts, a Copy button and a .txt download saved as extracted-text.txt
- No account and no sign-up; the image stays on your device
How to use image to text (ocr)
- “Drop an image here” or click “Choose image”Drag a screenshot, scan or photo onto the drop area, or click it to pick a file. JPG, PNG and WebP images work, and reading starts straight away in English.
- Under “Languages”, select every language in the imageTap English, Arabic, French, Spanish or Hindi; several can be on at once, and one always stays selected. The first use of a language downloads its data.
- Click “Extract text”Changing languages does not re-read the image by itself, so press “Extract text” to read it again with the new selection. The progress line runs from “Loading engine” to “Reading text”.
- Check the text, then “Copy” or “Download .txt”The box under “Extracted text” is editable, so fix any misread characters first. “New image” starts again with another picture.
FAQ
Is my image uploaded to a server?
No. Tesseract runs in a Web Worker inside your tab, and the image is handed to it straight from your device. The page downloads only the OCR engine and the language data, both from brevdigital.com itself, and neither request includes your picture.
Can it read Arabic text from an image?
Yes. Turn on Arabic under Languages, together with English if the image has both, and press Extract text. The Arabic data is a 1.7 MB download the first time. With Arabic selected, the result box switches to automatic text direction so right-to-left text displays the right way round.
Which languages does it support, and does it detect them?
English, Arabic, French, Spanish and Hindi. It does not detect the language: it reads with the languages you have selected, English by default, so switch on each language that appears in the image. Each extra language adds data to load and makes the read slower.
Which image formats can I use?
Screenshots, scans and photos saved as JPG, PNG or WebP. The picker only lists image files, and a dropped file that is not an image is ignored. Photos in Apple's High Efficiency Image format are not supported by the engine, so save them as JPG first. For a PDF, use the PDF editor, which has its own OCR.
Why is the first extraction slower than the next?
The first read in a language downloads that language's recognition data, from 1.4 MB for Hindi to 10.9 MB for English. Tesseract then keeps it in your browser's storage, so later reads load it from there instead. Each read still starts a fresh engine and closes it afterwards, which is why “Loading engine” appears every time.
Why is the text wrong or missing?
OCR depends on how clearly the characters show. If the box says “No text found — try another image or add its language.”, either nothing was recognised or the image's language is not selected. A sharper, straight-on image cropped close to the text usually reads better, and anything misread can be corrected in the box before you copy it.
Can I get the result as a Word file or a searchable PDF?
No, the output is plain text. Copy it into any document, or use Download .txt to save extracted-text.txt.