🧳 AI Trip Planner (Multi-Agent System)

An interactive Streamlit web app that uses multiple specialized AI agents to create comprehensive travel plans.

🔗 Live Demo

Features

• Interactive UI - Input your destination, trip duration, budget, and special preferences
• Multi-Agent Architecture - Four specialized AI agents work together to create your perfect trip

AI Agents

Agent	Role
🧠 Planner Agent	Creates day-by-day itineraries
💰 Budget Agent	Estimates and tracks costs
🍣 Local Guide Agent	Recommends food & local tips
🔍 Research Agent	Fetches current travel updates (used sparingly)
✈️ Travel Agent	Orchestrates all agents and produces the final plan

Installation

1. Clone the repository:

   git clone https://github.com/alfredang/trip-advisor.git
   cd trip-advisor

2. Install dependencies:

   pip install -r requirements.txt

3. Create a .env file with your API keys:

   OPENROUTER_API_KEY=your_openrouter_api_key
   TAVILY_API_KEY=your_tavily_api_key
   

4. Run the app:

   streamlit run trip_planner_openrouter.py

Usage

1. Enter your destination (default: Tokyo)
2. Set the number of days (default: 5)
3. Specify your budget in USD (default: $2,000)
4. Add any special preferences (optional)
5. Click "Generate Trip Plan"

The app will display your personalized trip plan with:

• 📋 Day-by-day itinerary
• 💰 Budget breakdown
• 🍣 Local recommendations and tips
• 🔍 Current travel updates (when available)
• 📥 Download option for your trip plan

Tech Stack

• Streamlit - Web interface
• OpenAI Agents SDK - Multi-agent orchestration
• OpenRouter - LLM API gateway
• Gemini 2.5 Flash Lite - LLM backend
• Tavily - Real-time web search
• Pydantic - Data validation

License

MIT