Build a Multi-Agent Travel Planning System using LangGraph + MCP

This project extends the Multi-Agent Travel Planning System built in Part 1 by integrating MCP (Model Context Protocol) servers for real-time flight and weather data.

Part 1 of This Project

GitHub Repository:
https://github.com/codewithaarohi/AI-Travel-Planning-System-using-LangGraph

Video Tutorial:
Build a Real-World Multi-Agent AI System using LangGraph | Multi-Agent AI + Memory + APIs
https://youtu.be/ctHby5vhDqg

---

Requirements

APIs

• Groq API: https://console.groq.com
• Tavily API: https://www.tavily.com/
• AviationStack API: https://aviationstack.com/
• OpenWeatherMap API: https://openweathermap.org/

Tools

• PostgreSQL: https://www.postgresql.org/download/
• Tavily MCP Server: https://docs.tavily.com/documentation/mcp

---

Step 1: Create Python Environment

python -m venv langgraph_env3

Activate:

langgraph_env3\Scripts\activate

---

Step 2: Install Dependencies

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

---

Step 3: Install PostgreSQL

Download PostgreSQL:

https://www.postgresql.org/download/

Important: Note your PostgreSQL password and port number during installation.

---

Step 4: Create Database

CREATE DATABASE langgraph_memory_demo;

---

Step 5: Setup .env File

Create a .env file:

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

---

Step 6: Get API Keys

• Groq: https://console.groq.com
• Tavily: https://tavily.com
• AviationStack: https://aviationstack.com
• OpenWeatherMap: https://openweathermap.org/

---

Setup AviationStack MCP Server (Local MCP Server)

Repository:

https://github.com/Pradumnasaraf/aviationstack-mcp

Open PowerShell:

E:

cd E:\Multi_agent_system_with_MCP

Clone repository:

git clone https://github.com/Pradumnasaraf/aviationstack-mcp.git

cd aviationstack-mcp

Install UV

Check:

uv --version

Install:

pip install uv

If installation fails:

powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"

Create .env File

AVIATION_STACK_API_KEY=your_api_key_here

Install Dependencies

uv sync

This will:

• Create .venv
• Install dependencies
• Install AviationStack MCP package

Activate Environment

.venv\Scripts\activate

Start MCP Server

uv run -m aviationstack_mcp mcp run

or

python -m aviationstack_mcp mcp run

The server will remain running and wait for MCP requests.

Stop Server

CTRL + C

---

Setup Weather MCP Server

Get API key:

https://openweathermap.org/

Add the API key to your .env file.

Install dependencies:

pip install mcp requests

---

Run the Application

Terminal Version

python main.py

Streamlit Web App

Copy frontend.py from the Part 1 repository into this project.

Run:

streamlit run frontend.py

---

Example Prompt

Plan a complete 7 days Japan trip including flights, hotels and sightseeing under 2 lakhs.

---

Features

• Multi-Agent Architecture using LangGraph
• PostgreSQL Memory
• Tavily Search Integration
• AviationStack MCP Integration
• Weather MCP Integration
• Streamlit Web App
• Real-Time Travel Planning