README.md•1.6 kB
# Multi-Agent Tools Platform
This project provides a modular, production-ready agentic system for advanced math, research, weather, and summarization tasks, using FastMCP, LangChain, and FastAPI.
## Features
- **Five core tools**: Math, Search, Weather, Wikipedia, Summarizer (`tools_server.py`)
- **Three specialized agents**: Research, Math, Meteo (`agents.py`)
- **HTTP API endpoints**: FastAPI server exposes each agent (`api_servers.py`)
- **Unified MCP toolbox**: Wraps APIs and provides a smart supervisor tool (`agent_tools.py`)
## Setup
### 1. Clone & Install
```bash
cd C:/Users/kingr/CascadeProjects
python -m venv venv
venv\Scripts\activate
pip install -r requirements.txt
```
### 2. Environment Variables
Create a `.env` file or set these in your shell:
```
AZURE_DEPLOYMENT=your-azure-deployment-name
OPENAI_API_VERSION=2023-05-15
```
### 3. Run Components (in separate terminals)
```bash
python tools_server.py
python api_servers.py
python agent_tools.py
```
- The FastAPI server runs on [http://127.0.0.1:8000](http://127.0.0.1:8000)
- Use an MCP client to interact with the `supervisor` tool in `agent_tools.py`
## File Structure
- `tools_server.py` – Five tools, FastMCP server
- `agents.py` – Three ReAct agents, using LangChain and FastMCP
- `api_servers.py` – FastAPI server, exposes each agent
- `agent_tools.py` – MCP toolbox, API wrappers, supervisor
- `requirements.txt` – All dependencies
- `.env.example` – Example environment file
## Testing
You can add tests using `pytest`. Example test files can be placed in a `tests/` directory.
---
## License
MIT