MCP Automations

A production-grade Model Context Protocol server in Python — four LLM-callable tools, two transports, deployed two different ways.

# 30-second proof the server is up:
curl -X POST https://mcp-automations.fly.dev/mcp \
  -H 'Content-Type: application/json' \
  -H 'Accept: application/json, text/event-stream' \
  -d '{"jsonrpc":"2.0","method":"initialize","id":1,
       "params":{"protocolVersion":"2024-11-05",
                 "capabilities":{},
                 "clientInfo":{"name":"curl","version":"1"}}}'

What this is

Most "AI engineer" portfolio projects are applications (a RAG chatbot, an agent that does research). This project is the layer underneath — the typed tools an LLM can call and the transport plumbing that exposes them. MCP is the emerging standard for LLM tool use (~97M monthly SDK downloads as of early 2026); building one — not just consuming one — is the rare skill.

For a deeper look at the design decisions — why two transports, how cost telemetry works, the exception hierarchy, what I'd do differently — see WRITEUP.md.

Tools

Plus one MCP resource (automations://catalog) and one MCP prompt (daily_brief) — using all three MCP primitives, not just tools.

Every tool returns a typed Pydantic model with per-call token usage and dollar cost attached. Cheap tasks route to Haiku 4.5 ($1/M in, $5/M out), writing-heavy tasks to Sonnet 4.6 ($3/M in, $15/M out).

Connect Claude Desktop to this server

Add one of these to ~/Library/Application Support/Claude/claude_desktop_config.json (Mac) or %APPDATA%/Claude/claude_desktop_config.json (Windows), then restart Claude Desktop.

Option A — local stdio (no network, runs the server as a subprocess):

{
  "mcpServers": {
    "mcp-automations": {
      "command": "python",
      "args": ["-m", "mcp_server.server"],
      "cwd": "/absolute/path/to/mcp-automations",
      "env": {
        "ANTHROPIC_API_KEY": "sk-ant-...",
        "TAVILY_API_KEY": "tvly-..."
      }
    }
  }
}

Option B — remote HTTP (talks to the live Fly server, no local setup):

{
  "mcpServers": {
    "mcp-automations": {
      "url": "https://mcp-automations.fly.dev/mcp",
      "transport": "http"
    }
  }
}

Then ask Claude something like "summarize https://www.paulgraham.com/greatwork.html in 3 bullets" — it'll call summarize_url automatically.

Run locally

pip install -r requirements.txt

# Stdio (for Claude Desktop):
python -m mcp_server.server

# HTTP server (defaults to 0.0.0.0:8765):
MCP_TRANSPORT=http python -m mcp_server.server

# Streamlit playground:
streamlit run playground/app.py

Requires Python 3.13. Needs ANTHROPIC_API_KEY and TAVILY_API_KEY in .env (see .env.example).

Architecture

┌────────────────┐     stdio      ┌──────────────────────┐
│ Claude Desktop ├───────────────►│                      │
└────────────────┘                │                      │
                                  │   mcp_server/        │
┌────────────────┐    HTTP/JSON   │   server.py          │
│ Remote client  ├───────────────►│   (FastMCP)          │
└────────────────┘   (Fly.io)     │                      │
              

... [View full README on GitHub](https://github.com/wzltmp/mcp-automations#readme)