Multi-Agent System · Open Source

Plan Your Perfect Trip
with AI Agents

VoyageAI uses 6 specialized AI agents powered by LangGraph to search flights, find hotels, discover activities, check weather, optimize budgets, and compile beautiful itineraries. 5 agents run instantly with zero API calls. Only the final compilation uses a free LLM.

View on GitHub See Live Demo
LangGraph LangChain Langfuse Free LLM Support Streamlit Open-Meteo API Zero-Cost Tools

6 Specialized AI Agents

Each agent is an expert in its domain, using dedicated tools and prompts to deliver the best results.

🌤️

Weather Agent

Fetches real weather forecasts using the free Open-Meteo API. Provides daily conditions, temperature ranges, and activity recommendations based on weather.

Open-Meteo API Geocoding Climate Analysis
✈️

Flight Agent

Searches for optimal flights with realistic pricing based on route distance and airline data. Provides booking links to Google Flights, Skyscanner, and Kayak.

Route Pricing Airline Database Price Comparison
🏨

Hotel Agent

Finds real hotels across budget tiers with amenities, ratings, and direct booking links to Booking.com, Expedia, Google Hotels, and more.

Hotel Database Price Comparison Review Analysis
🎯

Activity Agent

Discovers curated local experiences, attractions, and restaurant recommendations. Prioritizes authentic experiences with real costs and locations.

Curated Database Category Filter Restaurant Finder
💰

Budget Agent

Optimizes budget allocation based on real city cost data. Provides currency information, exchange tips, and smart spending recommendations.

Cost Database Budget Optimizer Currency Info
📋

Itinerary Agent

Uses an LLM to compile all agent results into a comprehensive day-by-day travel plan with weather-aware scheduling, booking links, and budget tracking. Falls back to structured output if no LLM is configured.

LLM Compilation No-LLM Fallback Day-by-Day Plan

Multi-Agent Architecture

LangGraph-powered sequential pipeline. 5 agents call tools directly (no LLM, instant). Only the final agent uses an LLM.

Weather Agent (direct tools - no LLM)
↓
Flight Agent
Hotel Agent
Activity Agent
All use direct tool calls — zero LLM cost, instant results
↓
Budget Agent (uses flight/hotel data)
↓
Itinerary Agent (LLM compilation — free via Groq/Gemini)
# LangGraph Workflow — Direct tool calls (no LLM for data agents)
from langgraph.graph import StateGraph, END

workflow = StateGraph(AgentState)

# Each node calls tools directly — fast, free, reliable
workflow.add_node("weather_agent", weather_node) # calls get_weather_forecast directly
workflow.add_node("flight_agent", flight_node) # calls search_flights directly
workflow.add_node("hotel_agent", hotel_node) # calls search_hotels directly
workflow.add_node("activity_agent", activity_node)
workflow.add_node("budget_agent", budget_node)
workflow.add_node("itinerary_agent", compile_itinerary_node) # only this uses LLM

# Sequential pipeline
workflow.add_edge("__start__", "weather_agent")
workflow.add_edge("weather_agent", "flight_agent")
# ... weather → flight → hotel → activity → budget → itinerary → END

Live Demo

Watch the multi-agent system plan a trip in real-time.

VoyageAI Terminal
$ python3 main.py --destination "Tokyo" --budget 3000 --dates "2025-06-01 to 2025-06-07"

Try It Yourself

Pick a destination and watch 6 AI agents build your itinerary

Why VoyageAI?

Built with the latest AI orchestration frameworks for intelligent, reliable travel planning.

🔗

Real Booking Links

Every flight, hotel, and activity comes with direct links to Google Flights, Booking.com, Skyscanner, and more.

🌡️

Real Weather Data

Live weather forecasts from Open-Meteo API (completely free, no API key required) to optimize your daily plans.

🤖

True Multi-Agent

6 specialized agents with direct tool calls. 5 run instantly without LLM. Only itinerary compilation uses a free LLM.

🆓

100% Free to Run

Works with free LLM providers (Groq, Google Gemini). Even produces results with zero API keys via fallback mode.

💸

Smart Budgeting

Real cost-of-living data for 30+ cities. Budget optimization with priority-based allocation.

🎨

Beautiful UI

Sleek Streamlit interface with real-time agent status tracking and downloadable itineraries.

🗺️

30+ Destinations

Curated activity databases for popular destinations. Works with any city via intelligent fallbacks.

🔓

100% Open Source

MIT licensed. Fork it, customize it, deploy it. Supports Groq (free), Google Gemini (free), and OpenAI.

Get Started in 30 Seconds

# Clone and setup
git clone https://github.com/sugeerth/agentic_data.git
cd agentic_data
pip install -r requirements.txt

# Set up a FREE LLM (choose one)
export LLM_PROVIDER=groq
export GROQ_API_KEY=your_free_key # Get at console.groq.com

# Or use Google Gemini (also free)
# export LLM_PROVIDER=google
# export GOOGLE_API_KEY=your_free_key # Get at aistudio.google.com

# Launch the beautiful UI
streamlit run app.py

# Or use the CLI (works even without API keys!)
python3 main.py -d "Tokyo" -b 3000 --dates "2025-06-01 to 2025-06-07"