Anydoc
Feeding real-world office documents, Word files, PowerPoint decks, spreadsheets, PDFs, into an LLM pipeline or search index requires converting them to clean text first, and most general-purpose converters either mangle formatting or run slowly at scale. anydoc is built specifically for that conversion problem: it turns Word, PowerPoint, Excel, OpenDocument, RTF, EPUB, CSV, and PDF files into clean Markdown. It is written in Rust for performance, with bindings available for Node.js and Python so it can be dropped into existing pipelines in either language.
As of August 2026 the repository has 12,808 stars and 53 open issues, indicating mature, high-adoption usage.
This is placed in Assess. It targets a common, concrete need, converting mixed office document formats into clean Markdown for downstream LLM or search use, with a performance-focused implementation and broad language bindings. There is no first-person production use yet, which is the gate for Trial.
