LLM-Friendly Documentation (llms.txt)
Waldur documentation supports the llms.txt standard to provide optimized access for Large Language Models (LLMs) and AI assistants.
What is llms.txt?
The llms.txt standard provides a structured way for AI systems to understand and navigate documentation. Instead of crawling HTML pages, LLMs can access a curated, text-only version of the documentation that's optimized for their context windows.
Available Files
/llms.txt
A curated index of Waldur documentation with descriptions for each section. This file helps AI agents understand the documentation structure and decide which resources to fetch.
URL: https://docs.waldur.com/llms.txt
Use case: Quick navigation, understanding documentation structure, deciding which pages to read.
/llms-full.txt
A comprehensive file containing all documentation content in a single text file. This enables loading the complete documentation context in one request.
URL: https://docs.waldur.com/llms-full.txt
Use case: Loading complete documentation into AI context, comprehensive searches, full-context assistance.
Using with AI Assistants
Claude (Anthropic)
You can reference the llms.txt files directly in conversations:
1 2 | |
ChatGPT / GPT-4
With web browsing enabled, GPT can fetch and use the llms.txt files:
1 2 | |
IDE Integrations (Cursor, Windsurf)
Many AI-powered IDEs support llms.txt natively. Add the Waldur documentation URL to your project's AI context settings.
MCP (Model Context Protocol)
The Waldur MCP Server provides direct API access that complements the llms.txt files for interactive AI assistance.
Content Structure
The llms.txt file organizes documentation by topic:
| Section | Description |
|---|---|
| Getting Started | Introduction, quick start, terminology |
| Deployment | Architecture, Helm, Docker Compose |
| Configuration | MasterMind settings, features, billing |
| Identity Providers | Keycloak, LDAP, SAML integration |
| Cloud Providers | OpenStack, Azure, SLURM setup |
| API Integration | Authentication, permissions, SDK |
| Developer Guide | Core concepts, development guides |
| User Guide | End-user documentation |
| Optional | API reference, changelogs (large files) |
Generation
The llms.txt files are generated using a script that:
- Reads the curated
docs/llms.txtindex file - Generates
llms-full.txtby concatenating documentation in priority order - Cleans content by removing images, videos, and HTML comments
- Excludes auto-generated API reference files to keep size manageable
Running Locally
1 2 3 4 5 6 7 8 | |
Updating llms.txt
The docs/llms.txt file is manually curated. When adding significant new documentation:
- Edit
docs/llms.txtdirectly - Add new sections or links following the existing format
- Use descriptive text after the colon for each link
- Keep the "Optional" section for large or auto-generated content
File Format
The llms.txt format follows the specification at llmstxt.org:
1 2 3 4 5 6 7 8 9 10 11 12 | |
Excluded Content
The following are excluded from llms-full.txt to keep the file size reasonable:
- API Reference (
/api-reference/): 270+ auto-generated endpoint files - API Changelogs (
/integrator-guide/APIs/api-changes/): 50+ version diff files - Images and videos: Binary content not useful for LLMs
- CHANGELOG.md: Large file with detailed release history
These resources remain accessible via direct URL or the API reference links in llms.txt.
Token Estimates
| File | Approximate Size | Estimated Tokens |
|---|---|---|
| llms.txt | ~6 KB | ~1,500 |
| llms-full.txt | ~2.2 MB | ~560,000 |
Token counts vary based on documentation updates. The llms-full.txt file is large and may exceed some LLM context windows. For models with limited context, use the curated llms.txt to identify specific sections to fetch.
Further Reading
- llms.txt Specification
- Waldur MCP Server - Model Context Protocol integration
- API Documentation - REST API reference