How do I extract text from an image?
To get text out of an image, upload it to FixDoks's Image to Text tool and pick the language. The Tesseract OCR engine reads the letters entirely in your browser, so the image is never sent to a server. Edit the result, then copy it or download it as a .txt file.
Image to Text (OCR) at a glance
| Engine | Tesseract OCR |
|---|---|
| Languages | 22, including Hindi |
| Output | .txt file or copy |
| Processing | In your browser; your files are not uploaded |
| Price | Free, no watermark |
| Download | Free Google sign-in |
| Works on | Chrome, Edge, Safari and Firefox on Windows, Mac, Android and iPhone |
How to use Image to Text Converter (OCR)
- Select an image, drop it on the box, or paste a screenshot with Ctrl+V (Cmd+V on Mac).
- Choose the language of the text. For mixed text, leave Also read English words ticked.
- Click Extract text. The text appears next to the image, where you can edit it.
- Click Copy text or Download .txt.
Turn pictures of text into editable text
Optical character recognition (OCR) reads the letters in an image and turns them into real text you can copy, search and edit. Use it to grab a paragraph from a screenshot, copy details from a photo of a notice, digitise a printed page, or pull the text out of an image someone sent on WhatsApp.
FixDoks uses Tesseract, a widely used open source OCR engine, running inside your browser. The image is never uploaded, which matters when it contains personal details such as an address, a bill or an ID number.
Supported languages
English, Hindi, Bengali, Marathi, Tamil, Telugu, Gujarati, Kannada, Malayalam, Punjabi (Gurmukhi), Urdu, Nepali, Arabic, Spanish, French, German, Portuguese, Italian, Russian, Chinese (Simplified), Japanese and Korean. The language data for the one you pick is downloaded the first time (usually 1 to 3 MB) and kept in your browser for next time.
Many Indian documents mix a regional language with English. Pick the regional language and keep Also read English words ticked, and both scripts are read in one pass.
How to get accurate results
- Sharp and straight: hold the phone parallel to the page and make sure the text is in focus. Blurry or tilted photos cause most mistakes.
- Good contrast: dark text on a light background works best. Avoid shadows across the page.
- Crop to the text: remove pictures, logos and busy backgrounds around the text.
- Resolution: small screenshots are enlarged automatically before reading, which improves accuracy on tiny fonts.
- Right language: reading Hindi with the English model gives nonsense, so always pick the language of the text.
What OCR can and cannot read
Printed text in common fonts is read very well. Handwriting, decorative fonts, text on curved surfaces and very low resolution images are much harder and will need correcting. Tables are read line by line, so columns may run together. Always proofread important details like names, amounts and ID numbers.
Several images at once
Select several images, for example photos of each page of a letter, and the text from each is joined in order, separated by a line. The confidence score shown at the end is the engine's own estimate of how sure it is. Low scores usually mean the photo needs to be sharper.
Line breaks
With Keep line breaks on, the text follows the lines in the image, which suits addresses, lists and poems. Turn it off to join the lines of each paragraph, which is better for pasting into a document or email.