PDF Text Per Page pulls the readable text out of a PDF and presents it one page at a time, as a tidy list of labelled cards. Instead of dumping an entire document into a single wall of text, it lets you see, copy or download the text of any individual page, the whole file, or just the pages that match a search term. It is built for anyone who works with long PDFs — researchers quoting a specific page, accountants pulling figures from a statement, developers scraping structured reports, or students collecting passages for an essay.
How it works
When you drop a PDF in, the tool loads it with pdf.js, the same battle-tested engine browsers use to render PDFs, bundled directly into the page so there is no network request. It walks every page in order and reads each page’s text-content stream: a list of text runs, each tagged with its position on the page. Those runs are reconstructed into readable lines. In layout mode the runs are grouped by their vertical position and sorted top-to-bottom and left-to-right, so line breaks, tables and addresses keep their shape. In flow mode the runs are concatenated into one continuous stream that pastes cleanly into another document. An optional pass re-joins words that were hyphenated across a line break.
Each finished page becomes its own card with a word count, a character count, a Copy button and a per-page .txt download. A live search box filters the cards down to only the pages containing your term, with matches highlighted in place, and the bulk Copy all, Download .txt and Download .md actions respect that filter so you export exactly the pages you care about. Pages that contain only a scanned image are detected and clearly labelled rather than left mysteriously empty. Your layout, hyphen and blank-page preferences are remembered in this browser for next time.
Example
Say you have a 48-page annual report and you only need the cash-flow statement on page 31. Open the PDF, type cash flow in the filter, and the list collapses to just the pages mentioning that phrase — each highlighted so you can confirm it at a glance. Click Copy on page 31 and paste it straight into your spreadsheet notes. Or leave the filter empty, click Download .md, and you get a Markdown file where every page sits under its own ## Page N heading, ready to drop into a wiki or a language model. If page 12 turns out to be a scanned signature page, the tool tells you it has no selectable text instead of silently skipping it.
Everything happens on your own machine. The PDF is never uploaded, so even sensitive financial or legal documents stay entirely private.