🧠 MCP Interactive Chatbot with LangChain and Groq

This project is an AI-powered interactive chatbot built using the MCP (Model Context Protocol), LangChain, and Groq's fast LLMs. It enables natural language conversations that can also interact with services like browsers, Airbnb listings, and search engines using MCP.

🚀 Features

• ✅ Asynchronous interactive chat interface
• ✅ Memory-enabled conversational agent powered by LangChain
• ✅ Seamless integration with MCP services (Playwright, DuckDuckGo, Airbnb)
• ✅ Super-fast response via Groq API
• ✅ Simple terminal-based experience with memory management (clear, exit)
• ✅ Scalable setup with modular service integration

🧰 Technologies Used

• Python 3.10+
• mcp_use
• LangChain
• ChatGroq from langchain_groq
• asyncio for non-blocking interactions
• .env for managing secrets

📁 Project Structure

mcp_demo/ │ ├── app.py # Main chat interface ├── .env # Stores your GROQ API Key ├── browser_mcp.json # MCP services configuration ├── requirements.txt # Project dependencies

🛠️ Installation

1. Clone the Repository

git clone https://github.com/your-username/mcp-demo-chatbot
cd mcp_demo

2. Install Dependencies

pip install -r requirements.txt

3. Add Your GROQ API Key

Create a .env file in the root directory:

GROQ_API_KEY=your_groq_api_key_here

4. Set Up MCP Services

Create a file named browser_mcp.json:

{
  "mcpServers": {
    "playwright": {
      "command": "npx",
      "args": ["@playwright/mcp@latest"]
    },
    "airbnb": {
      "command": "npx",
      "args": ["-y", "@openbnb/mcp-server-airbnb"]
    },
    "duckduckgo-search": {
      "command": "npx",
      "args": ["-y", "duckduckgo-mcp-server"]
    }
  }
}

Make sure Node.js and npx are installed and accessible from your terminal.

▶️ Running the App

python app.py

Then start chatting! For example:

• What is the weather in New York?
• Search for the latest AI news
• Find me Airbnb rentals in Bangalore

You can also:

• Type clear to reset memory
• Type exit to end the session

⚠️ Troubleshooting

• Takes too long to respond?

  • Ensure MCP services aren't hanging.
  • Try restarting the MCP subprocesses or simplify your prompt.
  • Check internet connection and firewall settings.

• FileNotFoundError?

  • Confirm that browser_mcp.json is in the same directory as app.py.

• Still not working?

  • Print debug logs in app.py or reinstall dependencies.

🧪 Sample Output

User: What’s the weather in New York?
AI: Based on recent searches via DuckDuckGo, it's currently 15°C and cloudy in New York.

💻 MCP Services You Can Add

You can extend the chatbot with even more services like:

• YouTube Search
• PDF Readers
• Finance APIs
• Calendar APIs
• Or custom tools you write yourself!

Just add them inside browser_mcp.json and make sure they're callable via MCP.

📦 Example requirements.txt

python-dotenv
mcp_use
langchain
langchain_groq

📝 License

MIT License. See the LICENSE file for details.

🙌 Credits

• MCP Protocol
• LangChain
• Groq API
• DuckDuckGo Search MCP
• Playwright MCP

🌐 Author

Vishwas R
Feel free to connect on LinkedIn or explore other projects on GitHub

Let me know if you’d like to turn this into a GitHub Pages landing page or want badges and shield links added at the top!