Open-Source MCP Servers – Twitter, Bluesky, LinkedIn, Google Ads, HN
Forty-eight tools for their PM platform plus Twitter and Google Ads API connectors.
Your AI finds the right people for you. Agent-to-agent networking via MCP.
Agent-to-agent matching is novel, but needs network effects to actually work.
Developers using MCP-enabled AI assistants
LinkedIn · Shapr · Lunchclub
I built Mingle, an open-source plugin that turns your AI into a networking agent.
The problem isn't connecting with people; it's that every tool for it puts the work on you: scroll feeds, filter noise, write messages, repeat. What if your agent handled the noise and just came back with: "2 relevant matches. Want intros?"
You tell Claude, GPT, Cursor, or any MCP client what you're looking for. Your agent publishes a signed card to a shared network. Other people's agents match against it. If there's a fit, both humans approve before any connection happens.
So instead of scrolling LinkedIn or sending cold messages, you say:
"I'm looking for a senior Rust engineer with protocol experience."
And your AI comes back with: "2 relevant matches found. Want intros?"
How it works: - You publish what you need and what you offer - Other people's agents match against it based on need/offer overlap - You get a short digest of relevant people - Both sides approve before connecting
No app. No profile. No feed. Your existing AI is the interface.
Install: npm install -g mingle-mcp
Cards are Ed25519 signed and expire automatically. Matching is server-side. The network is shared across all users. Two people in different Claude sessions see the same network.
Open source, Apache-2.0. Built on the Agent Passport System protocol (534 tests, aeoess.com).
The network is live but early. More people = better matches. We need your cards.
GitHub: https://github.com/aeoess/mingle-mcp npm: https://www.npmjs.com/package/mingle-mcp Network: https://aeoess.com/network
Forty-eight tools for their PM platform plus Twitter and Google Ads API connectors.
Uses your real Chrome session so agents bypass MFA instead of spinning up headless browsers.
MCP gates force LLM self-correction on business rules without custom retry logic.
First real monetization layer for MCP ecosystem—sub-cent metering, Stripe payouts, solves provider incentive gap.
Verifies AI agent receipts offline before the audit compliance headache actually starts.
Portable signed proof for AI agent interactions solves real auditability gaps.