Flight Search Assistant (Powered by Gemini + MCP)

What is Flight Search Assistant (Powered by Gemini + MCP)?

Flight Search Assistant (Powered by Gemini + MCP) is a Streamlit-based web tool that uses Google’s Gemini AI model and the MCP (Model Context Protocol) Server to find real-time flight information through natural language queries. It is designed for users who want a conversational interface to search for flights without navigating complex booking sites.

How to use Flight Search Assistant (Powered by Gemini + MCP)?

Clone the repository, install dependencies with pip install -r requirements.txt, create a .env file with GEMINI_API_KEY and SERP_API_KEY, then run streamlit run flight_search_ui.py. Open the provided local URL in your browser and enter a natural language flight query, such as “Find flights from Atlanta to Las Vegas on 2025-05-05”.

Key features of Flight Search Assistant (Powered by Gemini + MCP)

• Natural language flight search queries
• Real‑time flight data via SERP API through MCP
• Gemini Pro 2.5 model for query understanding
• Interactive Streamlit user interface
• Cleanly formatted flight results
• Error handling and fallback messaging

Use cases of Flight Search Assistant (Powered by Gemini + MCP)

• Quickly find flights between two cities on a specific date using plain English.
• Compare flight options without manually toggling filters.
• Demonstrate how AI agents can use MCP servers to access live external APIs.
• Prototype a conversational travel assistant with minimal setup.

FAQ from Flight Search Assistant (Powered by Gemini + MCP)

What API keys are required to run the tool?

You need both a GEMINI_API_KEY (from Google AI Studio) and a SERP_API_KEY (from serpapi.com). These must be stored in a .env file.

How does the tool process a natural language flight query?

The user’s query is sent to Gemini 2.5, which interprets it and generates a function call. That call is executed via the MCP Server (mcp-flight-search) to fetch flight data from the SERP API, and the results are displayed in the Streamlit UI.

Is the user interface a command line or a web app?

It is a web app built with Streamlit. After running the command streamlit run flight_search_ui.py, you interact with it in a browser.

What model does Gemini use in this project?

The project uses Gemini Pro 2.5 for understanding queries and generating tool‑use function calls.

Can I search for flights without an internet connection?

No. The tool relies on the SERP API for live flight data, so a working internet connection and valid API keys are required.