通话记录MCP服务

Call Logger MCP Server

FastMCP server with:

• MCP endpoint on /mcp
• Web dashboard on /
• JSON calls API on /api/calls
• SQLite persistence in data/calls.db

Prerequisites

• Python 3.13+
• uv installed (pip install uv)

Run Locally

1. Install dependencies:

uv sync --frozen --no-dev

2. Start the server:

uv run python call_logger_server.py

3. Open:

• Dashboard: http://localhost:8000/
• Calls API: http://localhost:8000/api/calls
• MCP (Streamable HTTP): http://localhost:8000/mcp

Run With Docker

1. Build image:

docker build -t call-logger-mcp .

2. Run container:

docker run --rm -p 8000:8000 call-logger-mcp

3. Open:

• http://localhost:8000/

MCP Tool

log_call(caller_name, caller_number, callee_number, risk_level, reason)

• risk_level allowed values: low, medium, high

API

GET /api/calls?limit=200

• limit is optional, min 1, max 1000
• Response shape:

{
  "count": 1,
  "calls": [
    {
      "id": 1,
      "caller_name": "John Doe",
      "caller_number": "+421900000000",
      "callee_number": "+421911111111",
      "risk_level": "low",
      "reason": "Payment issue",
      "created_at": "2026-02-17T11:00:00.000000"
    }
  ]
}