Convert any office document to Markdown in 5 milliseconds

A high-performance Rust library makes it painless to turn messy office files into clean Markdown.

Firecrawl released anydoc, an open-source Rust library that converts PDFs, spreadsheets, slides, and Word files into clean Markdown. It racked up over 4,800 GitHub stars this week. Because it skips heavy ML models, median conversions take under 5 milliseconds. It ships with bindings for Python, Node.js, WebAssembly, and a zero-install CLI.

Why it matters: AI models process Markdown much better than proprietary office formats. anydoc parses different document types into one shared model, keeping tables, footnotes, and headers consistent across files. It also installs as an Agent Skill, letting tools like Cursor and Claude Code read local files directly.

Try it from your terminal with npx @firecrawl/anydoc report.docx or test the WebAssembly demo online. Text-based PDFs convert locally on your machine, though scanned pages require Firecrawl's hosted OCR API.

Feeding context to your LLM just got a whole lot faster.