Skip to main content
Ruune exposes a Model Context Protocol (MCP) server that lets AI agents — Claude, ChatGPT, or your own custom agents — query your captured context, retrieve transcripts, search across notes, and manage action items through natural conversation.

​
Setup

Add Ruune to your MCP client configuration. Here’s an example for Claude Desktop:
claude_desktop_config.json
{
  "mcpServers": {
    "ruune": {
      "url": "https://mcp.ruune.ai",
      "headers": {
        "Authorization": "Bearer ruune_sk_live_xxxxxxxxxxxxx"
      }
    }
  }
}
The MCP server uses the same API key as the REST API. Generate one from your Dashboard → Settings → API Keys.

​
What Can Agents Do?

Once connected, your AI agent can:
  • Search across all your notes and conversations semantically
  • Retrieve full transcripts with speaker diarization
  • List and manage action items from your recordings
  • Get highlights you marked during a recording
  • Answer natural language questions about your captured context

​
Example Conversation

You: What did Sarah say about the API launch timeline in our last meeting? Claude (via Ruune MCP): In your “Product Sync with Sarah” meeting on June 15th, Sarah said she thinks you should prioritize the API launch and aim for end of Q3. She also mentioned the docs need to be ready by July. You both agreed to ship V2 by July 1 — that was marked as a highlight during the recording.

Tools Reference

Full list of available MCP tools with parameters and descriptions.

Authentication

How to generate and manage API keys for MCP access.

​
Supported Clients

  • Claude Desktop
  • Claude Code
  • Cursor
  • OpenAI Agents SDK
  • Custom MCP Clients