使用说明
MCP Travel Assistant Chatbot
An interactive Streamlit-based AI Travel Assistant powered by LangChain, OpenAI, and multiple MCP (Modular Component Protocol) servers. It allows the chatbot to coordinate with specialized back-end agents for flight search, hotel booking, weather info, finance insights, and more — all via a single unified chat interface.
Features
- Multi-Agent Orchestration: Communicates with multiple MCP servers (Flight, Hotel, Weather, etc.)
- LLM-Powered Chat: Uses OpenAI models (e.g.,
gpt-5-nano) for reasoning and response generation - Asynchronous Processing: Handles multiple server calls efficiently using
asyncio - Streamlit Chat UI: User-friendly conversational interface
- Environment-Based Configuration: Keys and paths are securely managed via
.env
Project Structure
mcp-travel-agent/
├── backend/
│ └── servers/
│ ├── flight_server/
│ │ └── flight_server.py
│ ├── hotel_server/
│ │ └── hotel_server.py
│ ├── event_server/
│ │ └── event_server.py
│ ├── geocoder_server/
│ │ └── geocoder_server.py
│ ├── weather_server/
│ │ └── weather_server.py
│ └── finance_server/
│ └── finance_search_server.py
├── frontend/
│ ├── app.py # Streamlit app
│ ├── config.py # MCP server configuration
│ └── .env # Environment variables (excluded from Git)
├── .gitignore # Ignored files like .env, __pycache__, etc.
├── README.md # Project documentation
└── LICENSE # MIT License
Installation
1. Clone the repository
git clone https://github.com/sushilparjapati/tripmind.git
cd mcp-travel-agent/frontend
2. Create and activate a virtual environment for each server (backend) and frontend
uv sync
3. Create a .env file inside your frontend folder
OPENAI_API_KEY=your_openai_api_key_here
SERPAPI_KEY=your_serpapi_key_here
4. How to run
streamlit run app.py
