This project is a Real-World Multi-Agent AI System built using LangGraph.
The system uses 4 AI agents that work together to plan a complete trip automatically.
✈️ Flight Search Agent- 🏨 Hotel Search Agent
- 🗓️ Itinerary Planning Agent
- 🤖 Final Response Agent
- 🧠 Memory using PostgreSQL
- 🌐 Real-time API Integration
- 💻 Streamlit Web Interface
- LangGraph
- LangChain
- Groq
- Llama 3.3 70B
- PostgreSQL
- Streamlit
- Tavily API
- AviationStack API
Open the terminal inside the project folder and run:
python -m venv langgraph_env3
Now activate the environment:
langgraph_env3\Scripts\activate
Run the following command:
pip install langgraph langchain langchain-openai langchain-groq langchain-community langchain-tavily psycopg[binary] psycopg_pool python-dotenv tavily-python requests streamlit
pip install -U "psycopg[binary,pool]" langgraph-checkpoint-postgres
Download and install PostgreSQL: https://www.postgresql.org/download/
- PostgreSQL Password
- Port Number
You will need them later while creating the database connection string.
Open PostgreSQL and run:
CREATE DATABASE langgraph_memory_demo;
Create a .env file inside the project folder.
Add the following keys:
GROQ_API_KEY=your_groq_api_key
TAVILY_API_KEY=your_tavily_api_key
AVIATIONSTACK_API_KEY=your_aviationstack_api_key
DATABASE_URL=postgresql://postgres:postgres@localhost:5433/langgraph_memory_demo
python main.py
This will test the multi-agent system through the terminal.
streamlit run frontend.py
This will launch the Multi-Agent AI web application.
Plan a complete 7 days Japan trip including flights, hotels and sightseeing under 2 lakhs.
- Flight Agent searches flights
- Hotel Agent searches hotels
- Itinerary Agent creates travel plan
- Final Agent combines everything together
- PostgreSQL stores conversation memory