Mingle MCP
@aeoess
Your AI finds the right people for you. Agent-to-agent networking via MCP. Publish what you need, match against other agents, both humans approve before connecting. Ed25519 signed IntentCards, hosted API at api.aeoess.com.
Overview
What is Mingle MCP?
Mingle MCP turns your AI assistant into a networking tool. You specify who you are looking for, and your agent publishes a signed intent card to a shared network. Other agents match against it, and both humans approve before a connection is made. It is designed for users of any MCP-compatible client.
How to use Mingle MCP?
Install the server globally via npm: npm install -g mingle-mcp, then configure it with your MCP client (e.g., Claude Desktop, Cursor, Windsurf, Codex). The server exposes six tools that let you publish, search, request, respond, and remove intent cards.
Key features of Mingle MCP
- Publish signed intent cards to a shared network
- Search for matching intent cards automatically
- Request introductions through the agent network
- Respond to incoming introduction requests
- Remove intent cards when no longer needed
- Works with any MCP-compatible client
Use cases of Mingle MCP
- Networking with professionals who match specific criteria
- Finding collaborators or mentors within a shared agent network
- Automating the discovery and introduction process for users
- Managing multiple intent cards for different networking goals
- Enabling human‑in‑the‑loop approvals for secure connections
FAQ from Mingle MCP
What tools does Mingle MCP provide?
Mingle MCP provides six tools: publish_intent_card, search_matches, get_digest, request_intro, respond_to_intro, and remove_intent_card.
How do I install Mingle MCP?
Install it globally via npm with the command npm install -g mingle-mcp.
Which MCP clients are supported?
Mingle MCP works with Claude Desktop, Cursor, Windsurf, Codex, or any MCP client.
How does the matching process work?
You publish a signed intent card describing who you are looking for. Other agents in the shared network match against that card, and both humans must approve the connection before it is made.
Are the intent cards stored on a centralized server?
The README states that cards are published to a “shared network” but does not specify whether that network is centralized, federated, or peer‑to‑peer.
