🏖️ VacAIgent: AI-Powered Travel Planner

An intelligent travel planning application built with CrewAI that uses multiple AI agents to create personalized trip itineraries. Available as a web app (Streamlit), command-line interface (CLI), and REST API (FastAPI).

🌟 Features

• Multi-Agent AI System: Leverages CrewAI framework with specialized agents for city selection, local expertise, and travel concierge services
• Multiple Interfaces: Choose between web UI, CLI, or API integration
• Comprehensive Trip Planning: Includes weather analysis, cost estimation, attraction recommendations, and detailed itineraries
• Real-Time Data: Integrates with search engines and web scraping for up-to-date information
• Flexible LLM Support: Supports Gemini, Groq, OpenAI, and local Ollama models
• Budget Analysis: Provides detailed cost breakdowns and packing suggestions

🚀 Quick Start

Prerequisites

• Python 3.8+
• API keys for:
  • Google Gemini (default LLM)
  • Serper.dev (web search)
  • Browserless.io (web scraping)

Installation

1. Clone the repository:

git clone https://github.com/Monish-Nallagondalla/crewai_travel_planner.git
cd crewai_travel_planner

1. Install dependencies:

pip install -r requirements.txt

1. Set up environment variables: Create a .env file in the root directory:

GEMINI_API_KEY=your_gemini_api_key
SERPER_API_KEY=your_serper_api_key
BROWSERLESS_API_KEY=your_browserless_api_key

📖 Usage

Streamlit Web App

Run the interactive web interface:

streamlit run streamlit_app.py

Navigate to http://localhost:8501 and enter your trip details in the sidebar.

Command Line Interface

Plan a trip via CLI:

python cli_app.py \
  --origin "Bangalore, India" \
  --destination "Krabi, Thailand" \
  --start-date 2024-05-01 \
  --end-date 2024-05-10 \
  --interests "2 adults who love swimming, dancing, hiking, shopping, local food, water sports"

REST API

Start the FastAPI server:

uvicorn api_app:app --reload

The API will be available at http://localhost:8000.

API Endpoints

• GET / - Welcome message and documentation links
• POST /api/v1/plan-trip - Generate trip plan
• GET /api/v1/health - Health check

Example API Request

curl -X POST "http://localhost:8000/api/v1/plan-trip" \
  -H "Content-Type: application/json" \
  -d '{
    "origin": "Bangalore, India",
    "destination": "Krabi, Thailand",
    "start_date": "2024-05-01",
    "end_date": "2024-05-10",
    "interests": "2 adults who love swimming, dancing, hiking"
  }'

🏗️ Project Structure

crewai_travel_planner/
├── api_app.py              # FastAPI application
├── cli_app.py              # Command-line interface
├── streamlit_app.py        # Streamlit web application
├── trip_agents.py          # AI agent definitions
├── trip_tasks.py           # Task definitions for agents
├── tools/                  # Custom tools for agents
│   ├── browser_tools.py    # Web scraping tool
│   ├── calculator_tools.py # Calculation utilities
│   └── search_tools.py     # Web search tool
├── images/                 # Static assets
├── requirements.txt        # Python dependencies
├── flow_diagram.svg        # Architecture diagram
├── .gitignore
├── LICENSE
└── README.md

🤖 AI Agents

City Selection Expert

• Analyzes weather patterns, seasonal events, and travel costs
• Compares multiple destination options
• Provides detailed reports on chosen cities

Local Expert

• Gathers in-depth information about attractions and customs
• Discovers hidden gems and local recommendations
• Creates comprehensive city guides

Travel Concierge

• Develops detailed itineraries with daily plans
• Provides budget breakdowns and packing suggestions
• Recommends specific hotels, restaurants, and activities

⚙️ Configuration

Environment Variables

| Variable | Description | Required | |----------|-------------|----------| | GEMINI_API_KEY | Google Gemini API key | Yes | | SERPER_API_KEY | Serper.dev search API key | Yes | | BROWSERLESS_API_KEY | Browserless.io scraping API key | Yes |

Switching LLM Models

Edit the llm parameter in TripAgents class:

# For Groq
self.llm = LLM(model="groq/deepseek-r1-distill-llama-70b")

# For OpenAI
self.llm = ChatOpenAI(model_name="gpt-4")

# For Ollama (local)
self.llm = LLM(model="ollama/llama3.2", base_url="http://localhost:11434")

📋 Dependencies

Key dependencies include:

• crewai - Multi-agent framework
• streamlit - Web UI framework
• fastapi - REST API framework
• uvicorn - ASGI server
• langchain - LLM integration
• requests - HTTP client
• unstructured - Document processing
• pydantic - Data validation

See requirements.txt for complete list.

🔄 Workflow

graph TD
    A[User Input] --> B[City Selection Agent]
    B --> C[Local Expert Agent]
    C --> D[Travel Concierge Agent]
    D --> E[Trip Itinerary]

📝 API Documentation

Interactive API documentation available at:

• Swagger UI: http://localhost:8000/docs
• ReDoc: http://localhost:8000/redoc

🤝 Contributing

1. Fork the repository
2. Create a feature branch: git checkout -b feature/amazing-feature
3. Commit changes: git commit -m 'Add amazing feature'
4. Push to branch: git push origin feature/amazing-feature
5. Open a Pull Request

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

⚠️ Disclaimer

• This application uses external APIs which may incur costs
• Ensure you have proper API access and monitor usage
• Trip recommendations are AI-generated and should be verified
• Always check current travel advisories and requirements

🙏 Acknowledgments

• CrewAI for the multi-agent framework
• Streamlit for the web interface
• FastAPI for the API framework