Content
# MCP Chatbot with Multi-Server Integration
An interactive chatbot powered by the **Model Context Protocol (MCP)** that integrates multiple specialized servers for enhanced functionality. Built with LangChain, Groq LLM, and MCP Agent for conversational AI with tool-calling capabilities.
## Features
- 🤖 **Conversational AI** with memory-enabled chat history
- 🔧 **Multi-Tool Integration** via MCP servers:
- **Playwright**: Browser automation and web interaction
- **Airbnb**: Search and retrieve accommodation listings
- **DuckDuckGo Search**: Web search capabilities
- 📝 **Logging**: Comprehensive logging for debugging and monitoring
- 💬 **Interactive CLI**: Simple command-line interface with conversation management
- 🧠 **Memory**: Persistent conversation history within sessions
## Prerequisites
- Python 3.11+
- Node.js and npm (for MCP server dependencies)
- Groq API key
## Installation
1. **Clone the repository**:
```bash
git clone <repository-url>
cd chatbot_model_context_protocol
```
2. **Install dependencies**:
```bash
pip install -r requirements.txt
# or using uv
uv sync
```
3. **Set up environment variables**:
Create a `.env` file in the project root:
```env
GROQ_API_KEY=your_groq_api_key_here
```
## Configuration
### MCP Servers
The chatbot uses three MCP servers defined in `mcp_servers.json`:
- **Playwright**: Browser automation for web scraping and interaction
- **Airbnb**: Search for accommodations with filters
- **DuckDuckGo Search**: General web search capabilities
You can modify `mcp_servers.json` to add or remove servers as needed.
### Logging
Logs are saved to `logs/mcpchatbot.log`. The logger configuration can be adjusted in `logger.py`.
## Usage
Run the chatbot:
```bash
python app.py
```
### Commands
- **Type your query**: Ask questions or give instructions
- **`clear` or `cls`**: Clear conversation history
- **`exit`, `quit`, or `q`**: Exit the chatbot
### Example Queries
```
You: Search for hotels in New York for March 1, 2026
You: Find information about the Eiffel Tower
You: Open a browser and navigate to example.com
```
## Project Structure
```
chatbot_model_context_protocol/
├── app.py # Main chatbot application
├── logger.py # Logging configuration
├── mcp_servers.json # MCP server definitions
├── .env # Environment variables (not in git)
├── requirements.txt # Python dependencies
├── pyproject.toml # Project metadata
├── uv.lock # Dependency lock file
└── logs/ # Log files (not in git)
```
## Dependencies
- `langchain-groq`: Groq LLM integration
- `mcp-use`: MCP Agent and Client
- `python-dotenv`: Environment variable management
- `asyncio`: Asynchronous execution
## License
MIT
## Contributing
Contributions are welcome! Please open an issue or submit a pull request.
