Analysis updated 2026-08-13
Convert Word, Excel, PowerPoint, or PDF files into clean Markdown for an LLM to read
Add an Agent Skill so your AI coding agent can read office documents it encounters
Batch convert a folder of mixed document formats into one consistent Markdown output
Extract structured document content, including embedded images, from a file in your app
| firecrawl/anydoc | memvid/memvid | iii-hq/iii | |
|---|---|---|---|
| Stars | 15,518 | 15,486 | 15,623 |
| Language | Rust | Rust | Rust |
| Setup difficulty | easy | moderate | moderate |
| Complexity | 2/5 | 3/5 | 4/5 |
| Audience | developer | developer | developer |
Figures from each repo's GitHub metadata at analysis time.
anydoc is a document conversion library built by Firecrawl that turns office style files into clean Markdown text. It reads Word, PowerPoint, Excel, OpenDocument files, RTF, EPUB, CSV, and PDF, and converts any of them into the same consistent Markdown output, no matter which format went in. The core is written in Rust, and the project ships bindings so the same conversion works from Node.js, Python, the browser through WebAssembly, or Rust itself. The simplest way to try it is a command line tool you run through npx, pointing it at a file and getting Markdown printed to your screen or saved to a file. There is also a hosted demo page that runs the whole thing as WebAssembly in your browser, so files never actually leave your machine during the demo. For developers who want to call it from code, each language binding exposes functions to convert a file path or raw bytes straight to Markdown, or to stop one step earlier at a structured document object that also keeps embedded images and other assets available. The README highlights several design choices: every format is parsed into one shared internal document model before being turned into Markdown, so headings, tables, lists, footnotes, and similar structure behave the same way regardless of source format. The format itself is detected by looking at the actual file bytes rather than trusting the file extension, so mislabeled files still convert correctly. Because it is pure Rust with no machine learning models or external services involved, conversion is reported as very fast, with a median time under five milliseconds per document, and PDF text extraction works locally without needing a separate OCR service. The library is also packaged as an Agent Skill, so an AI coding agent can install it and gain the ability to read office documents on its own. The README includes a benchmark comparing anydoc against six other conversion tools across 100 real documents in fourteen formats, where anydoc reports the widest format coverage and the fastest median conversion time among the tools tested. The project is released under the MIT license.
A fast Rust library, with Node.js, Python, and browser bindings, that converts Word, Excel, PowerPoint, PDF, and other office files into clean Markdown.
Mainly Rust. The stack also includes Rust, Node.js, Python.
Use freely for any purpose, including commercial use, as long as you keep the copyright notice, under the MIT license.
Setup difficulty is rated easy, with roughly 5min to a first successful run.
Mainly developer.
This repo across BitVibe Labs
Verify against the repo before relying on details.