travel-booking-agent

# Travel Booking Agent — AI Context

## Overview
Multi-agent AI travel booking system using LangGraph, Groq LLM, and PostgreSQL.

## File Structure
- frontend.py — Streamlit UI (dark theme, agent pipeline, travel plan download)
- main.py — LangGraph graph with 5 agents (flight, hotel, itinerary, final, email)
- tools/flight_tool.py — AviationStack API wrapper for real-time flight data
- tools/tavily_tool.py — Tavily web search wrapper for hotel/accommodation search
- requirements.txt — Python dependencies

## Architecture
User query → Flight Agent (AviationStack) → Hotel Agent (Tavily) → Itinerary Agent (Groq) → Final Agent (Groq) → Email Agent (SMTP) → Plan delivered

## Key Tech
- LangGraph StateGraph for orchestration
- Groq llama-3.3-70b-versatile as the LLM
- PostgreSQL + PostgresSaver for conversation checkpointing
- Streamlit for the frontend
- AviationStack API for flight data
- Tavily API for hotel/hotel search

## Environment Variables
GROQ_API_KEY, TAVILY_API_KEY, AVIATIONSTACK_API_KEY, DATABASE_URL, EMAIL_ADDRESS, EMAIL_PASSWORD