How to get text from a photo
- Choose the language written in the image. For a page with both, pick English + Hindi.
- Drop a photo, choose a file, or paste a screenshot with Ctrl+V.
- Press Get text. The first time you use a language, its file downloads once, which can take a few seconds on mobile data.
- Read the result and fix any mistakes in the box, then copy it or save it as a .txt file.
What people use it for
- Copying a long notice, circular or WhatsApp forward that arrived as an image.
- Typing up printed notes or a textbook page without retyping every word.
- Pulling an address, phone number or reference number off a photo of a document.
- Getting the text of a screenshot to translate or search.
Getting better results
Text recognition (OCR) is very good at clear printed text and weaker with everything else. These make the biggest difference:
- Light and flatness. Shadows and curved book pages cause most errors. Lay the page flat in daylight.
- Straight on. Photograph from directly above. Slanted text is harder to read.
- Crop first. Remove pictures, logos and table edges with Crop Image, so only the text is left.
- Big enough. Each letter should be at least about 20 pixels tall. Zoom in or move closer rather than photographing a whole page from far away.
- Printed beats handwritten. Neat block letters can work. Joined handwriting usually does not.
Hindi and other Indian scripts
Devanagari is harder for machines than English. Letters join along the top line, and matras sit above, below and beside the main letter, so a small blur can change a word. Clear printed Hindi from books, newspapers and government letters reads well. Expect to correct a few words, especially conjunct letters.
If your document mixes English and Hindi, choose English + Hindi. Choosing only one makes the other language come out as nonsense.
Tamil, Telugu, Bengali, Gujarati, Punjabi, Kannada, Marathi and Urdu are also available from the list.
Reading text from a PDF
If the PDF was made on a computer, you can usually select and copy its text directly in any PDF viewer, which is more accurate. If it is a scan, turn the pages into images with PDF to JPG at Print quality, then read each one here.
Privacy
Your image is read inside your browser. It is not uploaded. The only download is the language file itself, which comes to your device from a public content network and is reused next time.
Questions people ask
Is my image uploaded to a server?
No. The text reader runs inside your browser. Only the language file is downloaded to your device, once.
Why are some words wrong?
Blurry photos, shadows, unusual fonts and handwriting cause errors. Always read the result before using it.
Can it read a whole PDF?
Convert the PDF pages to images with PDF to JPG, then read each image here.
Does it keep the layout?
It keeps line breaks but not columns, tables or fonts. Expect plain text.
Why is the first run slow?
The language file is downloaded the first time. After that it is much faster.