Introducing mkdocs-llmstxt-md: Making Documentation LLM-Friendly
A new MkDocs plugin that bridges the gap between traditional documentation and AI/LLM workflows by serving markdown files alongside HTML and generating LLM-optimized index files.
python
llm
mkdocs
Author
Nok Lam Chan
Published
July 31, 2025
I’m excited to announce the release of mkdocs-llmstxt-md, a new MkDocs plugin that bridges the gap between traditional documentation and AI/LLM workflows.
The Problem
As AI tools become more integrated into development workflows, there’s a growing need for documentation that’s easily consumable by Large Language Models (LLMs). While HTML documentation looks great for humans, LLMs work much better with structured markdown content. This creates a friction point: do you optimize for human readers or AI consumption?
The Solution
mkdocs-llmstxt-md solves this by making your MkDocs documentation work seamlessly for both audiences. The plugin automatically:
Serves original markdown files at .md URLs alongside your HTML pages
Generates llms.txt index files that provide LLMs with a structured overview of your documentation
Creates llms-full.txt files containing your complete documentation in a single, LLM-friendly format
Adds copy-to-markdown buttons so users can easily grab markdown content
Key Features
Smart Content Organization
The plugin uses a flexible section-based configuration that supports glob patterns:
plugins:-llms-txt:sections:"Getting Started":-index.md:"Main documentation"-quickstart.md:"Quick start guide""API Reference":-"api/*.md" # Automatically includes all API docs
LLM-Optimized Output
The generated llms.txt provides a clean, structured index that LLMs can quickly parse:
# My DocumentationBrief description of the project## Getting Started- [Documentation](https://example.com/index.md): Main documentation- [Quick Start](https://example.com/quickstart.md): Quick start guide## API Reference- [Overview](https://example.com/api/overview.md)- [Functions](https://example.com/api/functions.md)
Universal Compatibility
Since the plugin generates standard markdown files and uses simple HTTP serving, it works with any hosting platform and doesn’t require special server configuration.
Real-World Impact
This approach has immediate benefits for teams using AI coding assistants:
Faster context loading: LLMs can quickly ingest documentation via direct markdown URLs
No HTML parsing: Clean markdown content eliminates formatting noise
Single-file option: The llms-full.txt provides complete documentation in one request
Human-friendly: Original HTML documentation remains unchanged
Modern Release Process
The project showcases modern Python packaging practices with GitHub Actions and PyPI Trusted Publishing - no API tokens required, just secure OIDC authentication.
That’s it! Your next mkdocs build will generate LLM-friendly versions of your documentation.
Looking Forward
This plugin represents a broader trend toward making development tools more AI-friendly without sacrificing human usability. As AI becomes more integrated into our workflows, we need infrastructure that works seamlessly for both human and machine consumption.
Try it out and let me know how it works for your documentation needs!