Media
4 items
Description
Turn a PDF into text or Markdown on your own computer, scanned pages included (OCR). No upload, no page limit. Get the text out of a PDF as plain text or Markdown, including pages that are only scanned pictures. On a tab showing a PDF, click the button (or press Alt+Shift+P): the PDF opens in the reader beside it and its text appears, page by page. Or open the reader and drop a PDF file on it. Copy the result, or save it as a .txt or .md file. Reads on your computer The PDF engine (pdf.js) and the text recognition for scanned pages (Tesseract) are built into the extension. Your file is never uploaded. When you click the button, the extension requests the tab's address once more to read the PDF there (if the tab shows a web page instead, it stops reading as soon as the reply says so). That is the only request it ever makes. Scanned pages in 13 languages Scanned pages are read in English, Spanish, French, German, Portuguese, Italian, Russian, Chinese (Simplified and Traditional), Japanese, Arabic, Hebrew or Persian. It starts in the language you last picked, or else in your browser's language when it is one of them; pick another and the PDF is read again. Text that is already in the PDF comes out in any language. Scanned pages are straightened before they are read (tested up to 15 degrees either way), a two-column scan is read one column at a time, and running heads and page numbers are left out of scanned pages too. Markdown that keeps the structure Headings are worked out from the type sizes, lists stay lists, and lines broken across the page are joined back into paragraphs. Reads columns in order, rebuilds tables as Markdown tables, and leaves out running heads and page numbers (it tells you how many). No limits No page limit, no daily quota, no account. Asks for very little It can't see a tab until you click its button or press its shortcut, and then only that tab's address. It remembers one thing, the language you last picked for scanned pages, on your computer. No history, no account, no tracking. Good to know - Vertical Chinese and Japanese text is not read. - Tables are rebuilt from the text a PDF carries, not from scanned pages, and only when their columns are spaced apart. Images are not rebuilt. - A word that has its own hyphen and is broken at the end of a line ("self-" then "selected") can come out joined. - A PDF that needs a password, or whose owner forbids copying its text, is not read. - A file on your computer can be dropped on the reader; to read a local PDF tab directly, turn on "Allow access to file URLs" for the extension. - The download is about 26.5 MiB, because the PDF engine, the recognition engine and its 13 language models are inside it.