Image to Text (OCR)

Reads English, Korean, Spanish, Portuguese, French and German. Images stay on your device.

Table in a phone photo or a scan? → Photo or scanned table to Excel (slower, but splits cells better)

Loading the tool...

When to use it

Copy text from a screenshot you cannot select

If a chat or web page will not let you copy text, take a screenshot and paste it with Ctrl+V or "Paste from clipboard". Keep "Language in the image" on "Korean + English", press "Read text", then "Copy" the result.

Turn a printed page into a .txt file

Photograph the page straight on, press "Read text", fix any wrong characters directly in the result box, and choose "Download .txt". An estimated accuracy is shown, so if it is low, retake the photo with better light.

Read an English label or receipt

For an English-only label or receipt, set "Language in the image" to "English" so Korean letters are not guessed by mistake. After it finishes, check numbers and spellings once, as the page reminds you to.

How to use

  1. Choose or paste (Ctrl+V) an image.
  2. Pick the language to start reading.
  3. Edit and copy the text or save as .txt.

To turn a whole scanned PDF into a searchable, copyable PDF, use OCR PDF. It keeps the original look and only adds the text on top.

How it works

Text is recognized by Tesseract.js (version 7, WebAssembly) running in a Web Worker in your browser. The recognition data for the chosen language (about 1 to 4 MB per language) is downloaded from this site the first time and cached in IndexedDB, while your images never leave your device. The page must be opened over the web (not as a local file) because Web Workers need it. Your file is processed in this browser and is not uploaded. To check this yourself, watch the browser's DevTools Network tab while you use the tool (step-by-step guide). Also available as a free REST API for developers and AI agents (edge-processed, zero retention).

Also available as: OCR Online

FAQ

Does it read handwriting?

Neat handwriting sometimes works, but printed text is far more accurate.

Why is the first run slow?

The recognition data downloads once on first use.

Is my image uploaded?

No. Text recognition runs in your browser. Only the recognition data (about 1 to 4 MB per language) is downloaded from this site the first time; your images are never sent anywhere.

Which picture formats work?

PNG, JPG, WEBP, GIF and BMP. iPhone HEIC photos will not open, so take a screenshot or convert the photo to JPG first.

How can I get more accurate results?

Use a sharp, straight photo and pick the language that matches the picture. If you get a memory warning, crop to just the part you need. Printed text is read much better than handwriting.

Can I read several images at once or other languages?

One image at a time. Languages: Korean + English, English, Korean, Spanish, Portuguese, French or German. For a whole scanned PDF, use the scanned-PDF text recognition tool instead.

Last checked: October 8, 2026 · Dagochim team

Related tools