Skip to content

How to convert a PDF to Markdown for ChatGPT, Claude and Gemini

Updated 26 September 2026 · 6 min read

AI tools can read PDFs, but they do not always read them well. Headings get lost, tables turn into jumbled numbers, and page headers and footers get mixed into the text. Converting the PDF to Markdown first gives the AI clean, structured text, which usually means better answers.

Why Markdown works well for AI

A PDF stores text as pieces placed at positions on a page. It does not really know what is a heading, what is a list or which column comes first. Markdown is plain text with a few symbols that mark the structure:

  • # marks a heading
  • - marks a bullet point
  • | separates table columns

AI models were trained on huge amounts of Markdown, so they understand this structure well. Clean text also uses fewer tokens, the units AI tools count and charge by, so more of your document fits in one conversation.

How to convert your PDF

  1. Open the PDF to Markdown converter.
  2. Drop your PDF into the box, or click Choose PDF file. The file is read in your browser and is not uploaded.
  3. Keep the default settings for most documents. They detect headings from font size, detect bullet and numbered lists, and join lines that were broken by the page width.
  4. Check the result in the preview. Fix anything by hand in the editor.
  5. Copy the Markdown and paste it into ChatGPT, Claude or Gemini, or download it as a .md file.

What to check after converting

  • Headings: headings are detected by font size. If your PDF uses bold text at normal size for headings, add # in front of them yourself.
  • Tables: tables with clear columns are turned into Markdown tables. Tables with merged cells may come out as plain lines; the Markdown Table Generator can rebuild them.
  • Two-column papers: the converter reads the left column first, then the right, so sentences stay in order.
  • Scanned PDFs: if the PDF is a photo or scan of paper, there is no text to extract. You will see a warning. These need OCR (text recognition), which this site does not do yet.
  • Bangla and other Indic scripts: PDFs often store these letters in drawing order, so vowel signs can come out mixed up. If you have the original Word file, use the Word to Markdown converter instead: it gives clean Bangla text.

How long is my document for the AI?

Paste the Markdown into the AI Token Counter to see roughly how many tokens it uses for GPT, Claude and Gemini, and whether it fits in one message. For very long documents, the same tool can split the text into chunks.

Is it safe for confidential documents?

The conversion runs entirely on your device, so the PDF never reaches our servers. Whether you then paste the text into an AI tool is your decision; check your organisation's policy on sharing documents with AI services.

Tools used in this guide

More guides