Extract text from an image or scanned PDF

Drop pictures or a PDF, paste a screenshot (Ctrl+V), choose files, or take a photo. The text is read on your device; nothing is uploaded.

Language of the text

The first time you read, the engine (about 7 MB) and 3 MB of language data download from this site, then your browser keeps them. Pick every language that appears in the picture; they can be mixed.

  • 100% Free
  • No Sign-Up
  • No Credit Card
  • No Free Trial
  • No Pro Tier
  • No Watermarks

Nothing is uploaded. Just open it and start.

Extract text from any image, free

Free, no sign-up. Turn a photo, screenshot or scanned PDF into editable text — drop it, paste it, or take a photo, and the words are read on your device in the language you choose. Proofread, copy, or download as .txt, .docx, .csv or a searchable PDF. Nothing is uploaded.

Ten languages, on your device

English, German, French, Spanish, Portuguese, Italian, Chinese, Japanese, Arabic and Hindi, mixed if you like, each loaded only when picked.

Images, screenshots and scanned PDFs

Drop, paste, choose or snap. Each picture keeps its own result, so one bad file never wipes the rest.

Proofread, then export

Uncertain words are marked. Copy, or download .txt, .md, .docx, .csv or a searchable PDF.

Private, no upload

Nothing leaves your device, there is no account, and the engine is served from this site.

Related tools

Word & Character Counter

Count words, characters, sentences and reading time in any language, with keyword density, readability and limits for X, SMS and SEO.

Image Converter

Convert images between PNG, JPG, WebP, SVG and more — free, in your browser, nothing uploaded.

Image Editor

Resize, crop, rotate, compress, and convert images — entirely in your browser.

Image Compressor

Compress JPG, PNG, WebP and AVIF in batches — aim for a file size like 200 KB, nothing uploaded.

Frequently asked questions

How does image-to-text work?
Drop pictures or a scanned PDF, paste a screenshot with Ctrl+V, choose files, or take a photo. The text is recognised in your browser with an on-device OCR engine (Tesseract). Each picture gets its own editable box, with the words the reader was unsure about marked so you know what to proofread, and you can copy everything or download it as .txt, .md, .docx or .csv.
Is anything uploaded?
No. Pictures and PDFs never leave your device: recognition runs locally, there is no server and no account. The reading engine and language data are files served from this site, downloaded the first time you use them (about 7 MB for the engine, plus about 1 to 3 MB for each language) and then kept by your browser. The page tells you when they have been loaded before, in which case it should work offline for as long as your browser keeps its saved files.
Which languages can it read?
English, German, French, Spanish, Portuguese, Italian, Chinese (Simplified), Japanese, Arabic and Hindi. Pick every language that appears in the picture (up to three at once, for example English and German); each one downloads only when you pick it. Text in a language you did not pick comes out garbled, for instance German read as English turns ß into R3.
Which formats work?
Any picture your browser can show (JPG, PNG, WebP, GIF, BMP) and iPhone HEIC photos, which are converted in the browser first. Scanned PDFs work too: every page (up to 40) is drawn and read, and labelled Page 1, Page 2 and so on. A password-protected PDF has to be unlocked first.
Can I get a searchable PDF?
Yes, for Latin-script languages. The Searchable PDF button builds a new PDF from the pictures you read, with the recognised words laid over them as invisible text so the pages look like the scans but can be searched and selected. It uses the text as first read, not your later edits.
It read my picture badly. What can I do?
Open Adjust picture on that result: turn it, straighten a crooked photo, cut away everything but the text, raise the contrast, or switch on Photo of a page to even out shadows, then Read again. Fill the frame with the text, shoot straight on, and use even light. Upside-down and sideways pictures are turned automatically by trying each way up and keeping the one the reader is surest about, and you can also turn them by hand. Handwriting is hit-and-miss.
How do I get tables or receipts into a spreadsheet?
Use the .csv download: each line becomes a row and wide gaps between words become columns. It works best on clean tables and receipts; check the columns after opening it. Join broken lines turns a wrapped paragraph back into one line and mends words split by a hyphen at the end of a line.

Spotted a bug or have a suggestion?

Found something broken, have an idea, or just want to say thanks about any of our tools? Every message reaches a real person.