Drop a PDF, get clean Markdown back. Free, in-browser, no sign-up. Headings, lists and structure are preserved so the output is ready for an LLM, a wiki, or a git diff.
PDFs lock content into a layout that's hard to search, hard to diff, and hard for an LLM to read. Markdown is plain text — every editor, every chat model, every static site generator understands it natively. Converting a PDF to Markdown gives you a file you can paste into Claude, ChatGPT, Notion, Obsidian or GitHub and actually use.
t0md extracts text, headings and list structure from the PDF and emits clean Markdown. Bullet and numbered lists, headings (H1–H6), inline emphasis and tables are preserved where the source PDF has real structure. Scanned image-only PDFs aren't supported yet — OCR is on the roadmap.
t0md exposes a remote MCP (Model Context Protocol) server. Add it once and any MCP-compatible agent can convert PDF to Markdown without leaving the conversation.
claude mcp add --transport http t0md https://t0md.com/mcpThen ask the agent: "convert /path/to/research-paper.pdf to markdown". The MCP server returns the converted Markdown plus a 10-minute download link. Full setup details on the MCP page.
Drop your PDF on this page (or click to choose a file). t0md converts it in seconds and gives you the Markdown to copy or download as a .md file. No sign-up, no upload limits beyond a 25 MB cap per file.
Markdown is the format LLMs, modern wikis, static site generators and note-taking apps all natively understand. A PDF locked behind a fixed layout becomes searchable, diffable and editable text the moment it's converted to .md.
We think so for the cost (free) and the integration story (MCP-ready, runs in-browser, no sign-up). For pure conversion accuracy on complex layouts, dedicated commercial tools may go further; t0md targets the everyday PDF-to-Markdown use case used by developers, researchers and AI agents.
Yes. t0md exposes an MCP server at https://t0md.com/mcp. Add it with `claude mcp add --transport http t0md https://t0md.com/mcp` and your agent can call convert_to_markdown on any PDF without leaving the conversation. Same setup works for Cursor, Windsurf, Hermes, OpenClaw and any MCP-compatible client.
No. PDFs are processed in memory and discarded as soon as conversion finishes. The browser converter never persists your file; the MCP download endpoint holds generated Markdown for at most 10 minutes before dropping it.