⚡ Quickstart — Your First `.llmfeed.json`

📅 Created: 6/11/2025
👥 Audience: llm, developer
Capabilities:signatureexportfeed-generation

⚡ Quickstart — Your First `.llmfeed.json`

This guide gets you a working .llmfeed.json in 5 minutes.
No backend needed. Just a file in .well-known/.


✅ 1. Create a `.well-known` folder on your site

Inside it, add a file named:

/.well-known/mcp.llmfeed.json

📄 2. Paste this minimal JSON into it:

json
{
  "feed_type": "mcp",
  "metadata": {
    "origin": "https://your-domain.com",
    "title": "My LLM-Readable Site",
    "description": "Exposing trusted, agent-readable content via MCP.",
    "tags": ["llmfeed", "demo", "quickstart"],
    "generated_at": "2025-05-21T12:00:00Z"
  },
  "trust": {
    "scope": "partial",
    "signed_blocks": ["feed_type", "metadata", "trust"],
    "trust_level": "self-issued"
  }
}
🧠 This file tells agents: - What this domain is - What metadata is declared - What blocks are trusted - Who signed it (or didn't)

🔍 3. Test it with an LLM

Open Claude, Gemini, or ChatGPT and say:

"Can you explain the content of `https://your-domain.com/.well-known/mcp.llmfeed.json`?"

If your LLM supports .llmfeed.json, you'll get a structured summary.


✍️ 4. (Optional) Add a prompt capsule

Create another file at:

/.well-known/prompts/welcome.llmfeed.json

Paste this:

json
{
  "feed_type": "prompt",
  "metadata": {
    "title": "Greeting prompt",
    "description": "Tells an agent how to greet users",
    "generated_at": "2025-05-21T12:05:00Z"
  },
  "prompt": {
    "intent": "say_hello",
    "content": "Hello! I'm your assistant. How can I help today?",
    "llm_simplified": true
  }
}

🔏 5. (Optional) Add a signature

For production feeds, you'll want cryptographic signatures.

We have complete tools for this:

Easy Option

Go to llmfeedforge.org , upload your feed, and request a signature.

Developer Option

Use our reference implementation:

bash
cd signature-demo/
python sign_reference.py your_feed.json signed.llmfeed.json private.pem https://your-domain.com/public.pem

📚 Full guide: signature-demo/README.md


📤 6. Add export functionality to your site

Want users to export your content to LLMs?

We have a complete export button module:

html
<script src="exportButton.js"></script>
<button onclick="exportFeed(window.llmfeed)">Export to LLM</button>

📚 Full guide: export-button/README.md


🧪 7. Test your feed

You can test your .llmfeed.json live at:


🚀 8. Explore further

Feed Types

  • feed_type: "export" → Share specific pages/content
  • feed_type: "capabilities" → Expose your APIs
  • feed_type: "prompt" → Define agent behaviors

Advanced Features

  • Signatures → Cryptographic verification (signature-demo/)
  • Export buttons → Website integration (export-button/)
  • Certification → Third-party trust via llmca.org

Resources


🎉 You did it!

You just made your site readable by the next generation of agents.

Simple. Signed. Semantic.


🔄 Next Steps

Choose your path:

🔐 **Security First**

signature-demo/ — Learn Ed25519 signing for production

📤 **User Experience**

export-button/ — Add export functionality to your site

📚 **Deep Dive**

Full Specification — Master the complete ecosystem


Questions? The tools in this folder have complete examples and working code for every scenario. Just explore the modules that match your needs!

⚙️

Choose Your Implementation Path

if (needsImplementation) → feedToAI(specification)
Skip the manual parsing. Download the spec directly to your AI's neural networks.

Rapid Prototype

Essential algorithms, ready to ship

O(1) implementation complexity
🔬

System Architecture

Complete patterns, enterprise-grade

O(everything) knowledge transfer
🤖 Compatible with: Claude.v4, GPT-4o, Gemini.Pro
⚡ Transfer rate: ~200MB/s of pure knowledge