Image to Text (OCR)
Copy the text out of photos, screenshots, scanned documents and PDFs. Drop in a picture, paste a screenshot with Ctrl+V, or take a photo with your phone, and get text you can copy, edit and save.
What “pictures never uploaded” means for this tool
When you open this page, your browser downloads the tool itself, the same way any web page loads. The first time you read a picture it also downloads the open-source text recognition engine (Tesseract) and the language you picked, about 15 MB, from this same website. From then on, your pictures are read inside your browser, on your own device. They are never sent to this website's server or to anyone else.
What is saved: the language you picked, in this browser. The engine is kept by your browser so it loads faster next time. Pictures and text are forgotten when you close the page.
What the server does see: like any website, the server that delivers this page sees basic connection details such as your IP address and which files were requested (for example the English language data). Never your pictures or the text in them.
Visit counting: this site counts visits with privacy-friendly analytics. No cookies, no ads, no tracking across other sites, and your IP address is not stored. It records which page you opened, your country, your device type, how long you stayed and when you download a result, never the file or anything in it.
JPG, PNG, WebP, GIF, BMP screenshots and photos, or scanned PDFs. Or paste a screenshot with Ctrl+V (⌘+V on a Mac).
How to use the Image to Text (OCR)
- Drop in a picture or PDF, paste a screenshot with Ctrl+V, or take a photo.
- Pick the language of the text.
- Wait a few seconds while the text is read.
- Copy the text or download it as a .txt file. Fix any mistakes right in the box.
Questions
How do I copy text from a picture?
Drop the picture in (or paste a screenshot with Ctrl+V). The text appears in a box next to it a few seconds later, ready to copy.
How accurate is it?
Very good on clear screenshots and straight, well-lit document photos. Blurry, tilted or low-light photos and fancy fonts give more mistakes; the confidence score tells you when to check closely.
Can it read handwriting?
Neat printed handwriting sometimes works, but it is made for typed and printed text. Joined-up handwriting usually comes out wrong.
Can it read a scanned PDF?
Yes. Drop the PDF in and every page is read in turn, up to 60 pages. To make the PDF itself searchable or turn it into Word, use the PDF tool.
Are my pictures uploaded?
No. The text is read by an open-source engine running in your own browser. Your pictures never leave your device.
Still stuck, or want a tool that isn't here? Ask on the Nebula Discord.