Firecrawl releases anydoc open-source Markdown conversion library

Firecrawl has released 'anydoc', an open-source Rust-based library designed to rapidly convert various document formats into Markdown. The tool aims to prepare data for easier consumption by artificial intelligence models.

Supported formats include Word (.doc, .docx, .docm), PowerPoint, Excel, OpenDocument, rich text, EPUB, CSV, and PDF. By converting non-PDF formats into a unified document model before generating Markdown, the library ensures consistent output for headings, tables, and footnotes regardless of the source version.

Built with Rust, the library does not rely on external services or machine learning models, allowing for high-speed processing with a median conversion time of under 5 milliseconds per document. It is available as a Rust library and can be used via Node.js, Python, or WebAssembly (WASM) in browsers. Note that it does not include OCR capabilities for scanned PDFs.