This MCP server implements a modular agent system built on Google's Agent Development Kit (ADK) and a FastMCP-based server. It includes a Coordinator Agent that routes requests to specialized agents: a Search Agent for web information retrieval, a Data Analysis Agent for handling CSV/Excel data and generating visualizations and reports, and a Code Generator Agent that can create source code from natural language descriptions. Clients interact via HTTP endpoints such as /ask to route to the best-suited agent, /search for direct web searches, /analyze-data for data-focused analysis, /generate-code for code generation, and /chat/completions for conversational interactions. To use it, start the server and send structured JSON requests to these endpoints; for example, post a query to /ask to obtain an analysis, a search result, or code generation depending on the content of your request.

Prerequisites:

• Python 3.9+
• Git
• Access to required API keys (OpenRouter and GROQ)

Install steps:

1. Clone the repository: git clone https://github.com/your-org/AI_Agent_with_MCP.git cd AI_Agent_with_MCP

2. Create and activate a virtual environment: python -m venv venv

   macOS/Linux

   source venv/bin/activate

   Windows

   venv\Scripts\activate

3. Install dependencies: pip install -r requirements.txt

4. Create a .env file (optional) and add API keys, or set environment variables directly: OPENROUTER_API_KEY=YOUR_KEY GROQ_API_KEY=YOUR_KEY

5. Run the MCP server: python server.py

The server will start on http://0.0.0.0:8080 by default.

Notes and tips:

• Ensure your OPENROUTER_API_KEY and GROQ_API_KEY are valid and accessible from the environment where you run the server.
• The Coordinator Agent determines the best specialized agent for a given request; structure requests clearly to improve routing accuracy (e.g., include file_path for data analysis or language and target output for code generation).
• Endpoints available: /ask (general routing), /search (web search), /analyze-data (data analysis), /generate-code (code generation), and /chat/completions (conversational).
• Review requests_log.txt for debugging and auditing requests/responses.
• If you add new agents or tools, update the coordinator routing logic so that the new capability can be invoked via /ask.
• If running behind a firewall or proxy, ensure the application can reach external web services for searching or data retrieval.