Extract Text from PDF

Every so often a PDF lands in your lap that is obviously text. You can see the words, you can even drag a cursor across them, and yet getting those words into a document turns into a fight with columns, headers and page furniture. I built this because I got tired of pasting a page into some other app just to read it back as plain text. Drop a PDF here and you get its text out, page by page, with a heading for every page number, a copy button beside each page, a copy all button for the lot, and a download as .txt when you want a file rather than a clipboard. Word and character counts sit at the top so you know the size of what you just pulled out.

It runs on pdf.js, the open source engine Firefox uses to draw PDFs on screen, loaded into your browser and pointed at the file you picked. Nothing is uploaded. The document never leaves your machine, and that is the whole reason this exists in this form, because the files people need text out of are contracts, statements, medical letters and filings. If you need to reshape the document itself rather than read it, the PDF splitter handles pages.

How to use

1. Drop your PDF onto the box, or press the button and choose it from your files.
2. Wait a moment while the pages are read on your device. Long documents take a few seconds.
3. Scroll the results. Each page appears under its own page number heading, in document order.
4. Press the copy button next to any single page to take just that page of text.
5. Use copy all when you want the whole document at once, page headings included.
6. Press download .txt to save everything as a plain text file you can open anywhere.
7. Check the word and character counts at the top if you need to know how much text you extracted.

Why use our extract text from pdf?

The page by page structure is the part people notice, and it is the part I got wrong first. My original version dumped everything into one enormous text box. That looked fine on a two page memo and was useless on a sixty page report where the thing you wanted was one clause on page 41. Now every page gets its own heading and its own copy button, so you can lift a single paragraph out of a long document without scrolling past everything else. The word and character counts matter more than I expected, because this text is usually heading somewhere with a limit attached.

Nothing is uploaded. Your browser opens the file, reads it on your device, and the text on screen was produced a few inches from your eyes instead of on somebody's server. I repeat that because upload based converters treat it as a footnote when it is the entire difference between the two approaches. The documents that get run through text extractors are the sensitive ones: leases, bank statements, settlement letters, employment agreements. Trading a copy of those to a free website in return for a .txt file is a bad deal, and here there is no deal to make.

There is one hard limit and I would rather you hear it from me than discover it on a deadline. This works on PDFs that contain real text, which covers almost everything exported from Word, Google Docs, accounting software, or a browser print to PDF. A scanned PDF is a different animal: it is photographs of paper wrapped in a PDF container, and there are no text characters inside it to extract, so pages come back empty. That is not a bug and no extractor of this kind can fix it. For scans the right door is the image to text tool, which uses recognition to read letters out of pixels.

Under the hood it is pdf.js, the PDF engine Mozilla maintains and Firefox ships. Using the same code path a browser uses to render a document means the text arrives in the order the file declares, and it means I am not babysitting a homegrown parser that quietly mangles ligatures and accented characters.

What I refuse to add: an account, an email field, a processing queue, and a clever summarize button that sends your document off somewhere. Extraction should be a ten second job that ends the moment you close the tab. There is no history, no dashboard, and nothing to delete afterwards, because nothing was ever stored.

Who is this tool for?

Researchers and students pull quotes out of papers and reports without retyping them, then run the result through the word counter to see whether the extract fits an assignment limit. Journalists lift wording out of a filing so they can quote it exactly. Anyone writing a summary starts by getting the source text into a place where it can be edited rather than admired.

In business the jobs are small and constant. Someone needs the payment terms out of a supplier contract to paste into an email. Someone is migrating old product sheets into a website and needs the copy, not the layout. Someone is sending a document to a translator who wants text, not a PDF.

Developers and analysts use it as step one of a scrape. Get the text, see what is really in the file, then decide whether the job is worth automating. It is also the fastest way to tell whether a PDF is text or a scan: if pages come back empty, you are dealing with images.

Frequently asked questions