OCR guide

Getting started with OCR

OCR reads the text in pictures and PDFs. The reading model runs in your browser, so your files are never uploaded.

Open OCR

Get the model once

The first time, the page says This tool needs a model. The model is PP-OCRv4 (about 15 MB). Nothing downloads until you ask for it.

1. Click Get more….
2. Read the licence and click Download.
3. When it is done, the file area appears. The model stays in your browser for next time.

Read the text

1. Drop images or PDFs on the box, or click Choose files. You can add several at once. To try it, click Try it with a sample receipt.
2. Click Read text. A progress bar shows how far it has come.
3. Each page shows with boxes around the lines it found. Untick Show boxes to hide them.
4. Use Previous, Next or the page list to move between pages and files.

Check, fix and save

The text of the page you are on is in the box on the right. You can edit it there before you save it.

Text (.txt), Markdown (.md) and Word (.docx) save all pages in order; the Word file starts each page on a new page. Searchable PDF keeps the picture of each page and puts invisible text on top, so you can search and select it. Copy all text puts it on the clipboard.

Tip: Read the text before you use it. Choose the language of the text under Language and model (85 languages, one model per script), and only horizontal lines are read. Prices on the right of a line can end up on their own lines.