Clean text from a PDF for AI tools
Pasting a PDF straight into a chatbot often brings broken lines, stray page numbers and lost structure. This converter reads the text layer of your PDF in your browser, joins lines back into paragraphs, turns larger text into Markdown headings, converts bullet and numbered lists, strips repeated headers and footers, and gives you a clean .md file that is ready for ChatGPT, Claude, a knowledge base or your notes. It also shows a rough token count so you can see whether the document fits a model's context window.
How it works
PDF files do not store headings or paragraphs, only positioned pieces of text with a font size. The tool groups text into lines by position, finds the most common font size as body text, and treats noticeably larger lines as headings, up to three levels. It merges lines that sit close together into paragraphs and repairs words split by a hyphen at the end of a line. Lines that repeat near the top or bottom of at least half the pages are treated as headers or footers and removed.
What it does not handle well
- Scanned PDFs and images of text have no text layer, so they need OCR first.
- Tables are not rebuilt. Their cell text comes out as plain lines, so check any table that matters.
- Multi-column layouts can interleave columns, and complex layouts such as magazines may need manual tidying.
- Mathematical formulas and special symbols depend on how the PDF encoded them.
Privacy
The conversion runs entirely in your browser. The PDF is never uploaded, so it is a reasonable way to prepare private documents. Always read the output before you rely on it: automatic structure detection is a best guess.