How to use Image to Text (OCR)
- 1Drop up to 30 images, or paste a screenshot with Ctrl+V.
- 2Choose the text language, and a second language if the text is mixed.
- 3Pick a text layout: Automatic for documents, Single block, or Scattered text for screenshots and signs. Keep Enhance image on.
- 4Click Extract text. Edit the result if needed, then click Copy text or Download .txt.
Getting the best results
- Use a sharp, straight, well lit image. Text should be at least 10 to 12 pixels tall.
- Choose the right language. Accented letters and non-Latin scripts need their own language data.
- Try Scattered text for screenshots, memes and photos of signs, and Single block for a paragraph or receipt.
- Enhance image converts to grayscale, stretches contrast and enlarges small images, which helps with screenshots and phone photos.
A confidence score is shown after reading. A low score means you should check the text carefully.
Privacy and how it works
The OCR runs on your device using Tesseract, an open source text recognition engine. Your images are never uploaded. The first time you use a language, the engine and that language's data file are downloaded from a public CDN and cached by your browser, so the first run is slower and needs an internet connection.
Limits
Handwriting, very stylized fonts, curved text and low resolution photos are hard for any OCR engine and may give poor results. Tables come out as plain text without columns. For PDFs, convert pages to images first or use a PDF text tool.
Frequently asked questions
How do I copy text from a screenshot?
Take the screenshot, press Ctrl+V (Cmd+V on Mac) on this page, choose the language and click Extract text. Then click Copy text.
Which languages are supported?
33 languages including English, Spanish, French, German, Portuguese, Russian, Arabic, Hindi, Chinese, Japanese and Korean. You can combine two languages for mixed text.
Why is the first run slow?
The OCR engine and the language data are downloaded once the first time you use a language. Later runs use the cached copy and are faster.
Can it read handwriting?
Only neat printed handwriting gives usable results. The engine is designed for printed text.
Are my images uploaded to a server?
No. The text is recognized in your browser. Only the OCR engine and language files are downloaded to your device.