Image to text
Extract text from an image: drop a photo, scan or screenshot and copy the words out of it, or download them as a .txt file. The OCR runs in your browser, so nothing is uploaded.
- 100% free
- Runs in your browser
- No uploads
- Runs on your deviceEvery tool works directly in your browser. No files are uploaded.
- Your privacy mattersYour files never leave your device, so there's nothing for us to store.
- No limitsUse any tool as much as you need. No daily caps, no waiting.
- Completely freeNo signup and no trial to cancel. Ads keep the tools free.
How to extract text from an image
- Pick the language. English is selected. Hindi, Spanish, French, German, Portuguese are also available.
- Add your images. Drop them on the box, choose them, or paste a screenshot with Ctrl+V. You can add several; they are read one after another.
- Wait for the text. The first run downloads the reader and the language data once, which takes a few seconds on a good connection. A bar shows how far each image has got.
- Copy or download. Fix any mistakes in the text box, then copy it or save it as a .txt file. With several images, Download all puts every result in one file.
How the text is read
This tool uses Tesseract, a free, open-source OCR (optical character recognition) engine first developed at HP and later by Google and volunteers. Its current version finds the lines of text in an image and reads each line with a neural network trained on printed text. Here it runs as WebAssembly in a background worker, so the page stays usable while it works.
Before reading, each image is turned the right way up from its camera data and drawn on white, so text on a transparent PNG doesn't vanish. Small images such as screenshots are enlarged two times and very large photos are reduced to 4000 px on the longest side, because the engine reads best when letters are a moderate size. The optional Clean up faint or dark photos step makes the image grey, stretches its contrast and turns light-on-dark text into dark-on-light.
What to expect from different images
| Image | Typical result |
|---|---|
| Screenshot of a document, web page or chat | Very good. Usually only a few words need fixing. |
| Flatbed scan of a printed page | Very good, especially at 300 DPI or more. |
| Phone photo of a book or letter, flat and in focus | Good. Curved lines near a book's spine cause more mistakes. |
| Dark-mode screenshot, light text on black | Good, and often better with the clean-up option. |
| Shop receipt | Mixed. Faded thermal print and creases lose characters; check every price. |
| Table or spreadsheet | The words are read, but the columns are lost. |
| Handwriting, signs at an angle, text over a photo | Poor. Expect missing or wrong words. |
Tips for better results
- Crop to just the text you need with Crop image. Logos, photos and page edges can be read as stray letters.
- Take the photo straight on, in even light, with the page filling the frame and no shadow from your phone.
- Turn sideways photos upright first; the reader expects lines running left to right. A quick quarter turn in Rotate image fixes it.
- Choose the language of the text. English models misread accents, and the Hindi model is needed for Devanagari.
- If the result is poor, tick the clean-up option and compare; it helps most with faded, grey or dark images.
Frequently asked questions
Are my images uploaded to a server?
No. The text reader runs inside your browser tab, in a background worker. The reader and its language data are downloaded from this site the first time you use them, and your images never leave your device.
Can it read handwriting?
Not well. The models are trained on printed and typed text. Neat block capitals sometimes come out readable, but joined-up handwriting usually turns into nonsense. For notes, typing them is often quicker than fixing the result.
How do I convert a scanned PDF to text?
Save the PDF's pages as images first with PDF to JPG, then drop those images here; you can add every page at once and download all the text as one .txt file. If the PDF was made from a Word file or a web page, it already contains real text, so PDF to Markdown gets it out faster and more accurately.
Why is some of the text wrong?
Blur, low light, a tilted page, small print and busy backgrounds all cause mistakes, and similar shapes get swapped: 0 and O, 1 and l, rn and m. Check names and numbers. A sharper, straighter photo taken in daylight helps more than any setting.
Which languages can it read?
English, Hindi, Spanish, French, German and Portuguese. Pick the language before you add images, or change it afterwards and the images are read again. Each language is a download of 2 MB or less the first time, and is then kept by your browser.
Does it keep the layout, columns or tables?
No. You get plain text, read top to bottom. Paragraphs and line breaks are mostly kept, but the cells of a table come out as lines of words separated by spaces, and two-column pages can mix their columns.