Docx

Pulls text out of Word documents using python-docx, which is exactly what you need when you're dealing with .docx files in a pipeline and don't want to open Word or mess with XML parsing yourself. Triggers automatically on .docx extensions or when you mention "word document" in your intent. The implementation is straightforward, extracting paragraph text without trying to preserve formatting or handle tables and images. It's AGPL licensed by axoviq.com, so keep that in mind if you're working on proprietary stuff. Does one thing cleanly, though you'll want something beefier if you need to preserve document structure or extract non-text elements.