AnyDoc is free and open source, from the team behind Firecrawl: https://github.com/firecrawl/anydoc
What it does
Turns Word, PowerPoint, Excel, OpenDocument, RTF, EPUB, CSV and PDF files into clean Markdown, keeping headings, tables and lists. Old .doc files too. On my Mac, a Word proposal with a pricing table came out clean, table intact, in about two seconds.
Why bother
Messy formatting wastes ChatGPT's effort and breaks tables.
Run it (pick one)
With Node: npx @firecrawl/anydoc yourfile.docx
With Python: pip install firecrawl-anydoc
As a Claude Code skill: npx skills add firecrawl/anydoc
Where to use it first
Old proposals, contracts and price sheets, before you paste them into ChatGPT or Claude.