awesome-llm-apps/mcp_ai_agents/ai_travel_planner_mcp_agent_team

🌍 AI Travel Planner — MCP Agent Team (Airbnb, Google Maps, Weather)

This project is a multi-agent travel planning assistant built using Agno, powered by multiple Model Context Protocol (MCP) tools. It integrates:

•⁠ ⁠🏠 Airbnb Listings
•⁠ ⁠🗺️ Google Maps for routing and places
•⁠ ⁠🌦️ Weather information via AccuWeather
•⁠ ⁠📅 (Optional) Google Calendar integration via Gumloop MCP

All handled through a simple terminal interface powered by Python + asyncio.

---

⚙️ Requirements

•⁠ ⁠Python 3.10+ •⁠ ⁠Node.js + ⁠ npx ⁠ •⁠ ⁠uvx for running MCP servers from Git •⁠ ⁠Internet access for MCP calls

---

📦 Installation

1.⁠ ⁠Clone the repo

⁠ bash git clone https://github.com/yourusername/ai-travel-planner.git cd ai-travel-planner  ⁠

2.⁠ ⁠Create Virtual Env

conda create -n env_name python=3.10 -y conda activate env_name pip install -r requirements.txt

3.⁠ ⁠Create env file

GOOGLE_MAPS_API_KEY=your_google_maps_api_key ACCUWEATHER_API_KEY=your_accuweather_api_key OPENAI_API_KEY=your_openai_api_key

🧠 Built With

-- Agno

-- FastMCP

-- OpenAI

-- Google Maps Platform

-- Airbnb MCP

-- Accuweather

---

📅 Calendar MCP Server (calendar_mcp.py)

This script is a dedicated MCP (Modular Cognitive Process) server that integrates with Google Calendar to create events via API calls. It works seamlessly with the rest of the multi-agent system built with Agno and FastMCP.

---

🛠 Features

• Creates Google Calendar events using the create_event tool
• Accepts event title, description, date/time, location, attendees, and reminders
• Uses a refresh token for persistent access (no re-authentication needed)

---

🌱 Setup

Add the following to your .env file in the root directory:

GOOGLE_CLIENT_ID=your_google_client_id
GOOGLE_CLIENT_SECRET=your_google_client_secret
GOOGLE_REFRESH_TOKEN=your_refresh_token