AI Integrations¶

SRv6.md is designed to be AI-friendly. Use our knowledge base as context for your favorite AI assistant to get accurate, up-to-date answers about SRv6.

Option 1: llms.txt (Any AI)¶

We provide standardized files that any LLM can consume:

How to Use¶

Simply paste one of these URLs in your conversation with any AI:

Read https://srv6.md/llms-full.txt and use it as context.
Then answer: How do I configure End.DT4 on Linux?

Works with Claude, ChatGPT, Gemini, and any AI that can fetch URLs.

Option 2: Claude Project¶

Upload the wiki as a Knowledge Base in a Claude Project for persistent context.

Steps¶

1. Go to claude.ai → Projects → New Project
2. Name it "SRv6 Expert" (or whatever you prefer)
3. Download the docs from GitHub
4. Upload all .md files from docs/ as Knowledge
5. Set the project instructions:

You are an SRv6 networking expert. Use the uploaded knowledge base
to answer questions about Segment Routing over IPv6, including
configurations, RFCs, and lab setups. Always cite the relevant
page when answering.

Now every conversation in that project has full SRv6 knowledge.

Option 3: MCP Server (Claude, Cursor, any MCP client)¶

The MCP (Model Context Protocol) server gives your AI assistant direct tool access to the SRv6.md knowledge base. It is free and hosted: nothing to install, just add the URL.

https://mcp.srv6.md/mcp

Available Tools¶

Setup for Claude Code¶

claude mcp add --transport http srv6 https://mcp.srv6.md/mcp

Setup for Claude Desktop, claude.ai, and other clients¶

Add a remote MCP server (custom connector) with the URL https://mcp.srv6.md/mcp. Clients that only support local servers can bridge it with mcp-remote:

{
  "mcpServers": {
    "srv6": {
      "command": "npx",
      "args": ["-y", "mcp-remote", "https://mcp.srv6.md/mcp"]
    }
  }
}

What You Can Ask¶

• "What SRv6 behaviors does the Linux kernel support?"
• "Show me the IOS-XR config for SRv6 L3VPN"
• "What's the difference between End.DX4 and End.DT4?"
• "Give me an FRR uSID locator template with prefix fcbb:bb00:2::/48"

For Developers¶

Build Your Own Integration¶

The content is available in multiple formats:

RAG Pipeline Example¶

# Example: Build a RAG pipeline with the wiki content
import requests

# Fetch all content
response = requests.get("https://srv6.md/llms-full.txt")
content = response.text

# Split into chunks, embed, and store in your vector DB
# Then query with your LLM of choice