Mario AI Portfolio

Dual-protocol AI portfolio server — an MCP server for structured tool access and an A2A agent ("Have You Met Mario?") for conversational discovery. Both serve the same data about my experience, skills, projects, and services.

Why This Exists

Traditional portfolios are static. This one lets potential clients interact with my portfolio using their own AI tools — ask specific questions, get structured answers, and discover what I can do for them. The medium demonstrates the skill being sold.

Architecture

mario-ai-portfolio/
├── src/
│   ├── server.py            # Combined ASGI dispatcher (production entry point)
│   ├── mcp_server.py        # FastMCP server — 7 tools
│   ├── a2a_server.py        # A2A agent — conversational interface
│   └── data/                # Shared content modules
│       ├── about.py
│       ├── skills.py
│       ├── services.py
│       ├── projects.py
│       ├── experience.py
│       └── contact.py
├── Dockerfile               # Combined production image
├── Dockerfile.mcp           # Standalone MCP server
├── Dockerfile.a2a           # Standalone A2A agent
└── render.yaml              # Render deployment config

MCP Server

7 tools exposed via Streamable HTTP:

Connect to the MCP Server

Add the following config to your MCP client of choice:

Claude Code — run in your terminal:

claude mcp add mario-portfolio --transport http https://<your-service>.onrender.com/mcp

Claude Desktop — add to claude_desktop_config.json:

{
  "mcpServers": {
    "mario-portfolio": {
      "url": "https://<your-service>.onrender.com/mcp"
    }
  }
}

Cursor / VS Code — add to .cursor/mcp.json or .vscode/mcp.json:

{
  "servers": {
    "mario-portfolio": {
      "url": "https://<your-service>.onrender.com/mcp"
    }
  }
}

Then just ask your AI assistant anything about Mario — skills, projects, services, availability.

A2A Agent — "Have You Met Mario?"

Conversational agent-to-agent interface powered by Llama 3.1 8B via Groq. Supports agent discovery via the A2A protocol.

• Agent Card: https://<your-service>.onrender.com/.well-known/agent.json
• Skills: portfolio_query, service_inquiry, availability_check

Interact with the A2A Agent

Discover the agent — fetch the Agent Card:

curl https://<your-service>.onrender.com/.well-known/agent.json

Send a message — via JSON-RPC 2.0:

curl https://<your-service>.onrender.com/ \
  -X POST \
  -H "Content-Type: application/json" \
  -d '{
    "jsonrpc": "2.0",
    "id": "1",
    "method": "message/send",
    "params": {
      "message": {
        "role": "user",
        "parts": [{"kind": "text", "text": "What projects has Mario built?"}],
        "messageId": "msg-001"
      }
    }
  }'

From Python — using the a2a-sdk client:

from a2a.client import A2AClient

async with A2AClient(url="https://<your-service>.onrender.com/") as client:
    card = await client.get_card()
    print(card.name)  # "Have You Met Mario? — AI Automation Engineer"

Any A2A-compatible agent or orchestrator can discover and interact with this agent automatically via the Agent Card endpoint.

Tech Stack

• MCP: FastMCP v3 — Python, Streamable HTTP
• A2A: a2a-sdk — Official SDK, JSON-RPC 2.0
• LLM: Llama 3.1 8B Instant via Groq
• Deployment: Docker, Render (free tier)

Run Locally

Requires GROQ_API_KEY in a .env file.

# Combined server (MCP + A2A on port 8000)
uv run python -m uvicorn src.server:app --reload

# Or run services independently:
uv run python -m uvicorn src.mcp_server:app --port 8000   # MCP only
uv run python -m uvicorn src.a2a_server:app --port 9000   # A2A only

Verify locally:

curl http://localhost:8000/health
curl http://localhost:8000/.well-known/agent.json

Deployment

Single Docker service deployed on Render free tier.

The service stays permanently warm via a UptimeRobot monitor pinging /health every 5 minutes (free plan).

Deploy your own

1. Fork this repo
2. Create a Render account (no credit card required)
3. New → Web Service → connect your fork — Render detects render.yaml automatically
4. Set environment variables:

• GROQ_API_KEY — your Groq API key
• AGENT_URL — the Render service URL (set after first deploy)

1. Set up a free UptimeRobot HTTP monitor on /health at 5-minute intervals

License

MIT