# Personal Ops Assistant (MCP)

An AI-powered personal operations assistant built using the **Model Context Protocol (MCP)**, **FastMCP**, **Google Gemini**, and **FastAPI**. 

It dynamically discovers and executes operational tools at runtime to search local notes, inspect GitHub repository activity, and summarize live web pages through a web interface.

---

## 🚀 Key Features

- **Model Context Protocol (MCP):** Connects a Gemini client agent to a local FastMCP tool server via `stdio`.
- **Dynamic Tool Discovery:** MCP tools are listed, converted to Gemini function declarations, and invoked dynamically at runtime.
- **Integrated Tools:**
  - `search_local_notes`: Searches local `.md` and `.txt` files for keywords and returns file content.
  - `github_recent_activity`: Fetches recent commit logs from any GitHub repository using the GitHub REST API.
  - `summarize_webpage`: Extracts and summarizes content from live URLs.
- **Web UI:** Modern chat interface with live MCP tool call inspection logs.

---

## 📁 Project Structure

```text
├── notes/              # Local markdown and text notes directory
├── static/
│   └── index.html      # Responsive chat interface
├── agent.py            # Gemini client & MCP dynamic tool orchestration loop
├── server.py           # FastMCP server exposing local tools over stdio
├── app.py              # FastAPI server & API endpoints
├── requirements.txt    # Project dependencies
└── README.md           # Project documentation