A production-ready AI travel assistant built with LangGraph that helps users find flights, book hotels, and manage travel plans through intelligent multi-agent interactions.

• Multi-Agent Architecture: Leverages different LLMs for specialized tasks
• Stateful Conversations: Maintains context across interactions
• Human-in-the-Loop: User approval for critical actions
• Email Integration: Automated travel plan delivery
• Real-time Data: Live flight and hotel information

• Python 3.11+
• Poetry

1. Clone the repository:

git clone [email protected]:zeroai/ai-travel-agent.git
cd ai-travel-agent

2. Set Python version (if using pyenv):

pyenv local 3.11.9

3. Install dependencies:

poetry install --sync
poetry shell

Create a .env file in the root directory:

OPENAI_API_KEY=your_openai_api_key
SERPAPI_API_KEY=your_serpapi_api_key
SENDGRID_API_KEY=your_sendgrid_api_key
LANGCHAIN_API_KEY=your_langchain_api_key
LANGCHAIN_TRACING_V2=true
LANGCHAIN_PROJECT=ai_travel_agent

Start the application:

streamlit run app.py

Example request:

"I want to travel to Amsterdam from Madrid from October 1st to 7th. Find me flights and 4-star hotels."

The system uses LangGraph to orchestrate multiple AI agents:

• Flight Finder: Searches and compares flight options
• Hotel Finder: Locates accommodation based on preferences
• Email Generator: Creates personalized travel summaries
• Coordinator: Manages workflow and user interactions

MIT License - see LICENSE for details.