Gemini With MCP Servers

Single AI chat platform that unifies esports and FPL data using MCP, letting fans get real-time schedules and stats through natural conversation, and easily extendable to more sports in the future.

Comment

Inspiration

We wanted to create a single conversational AI interface that eliminates the need to jump between multiple platforms for esports and fantasy sports information. With the rise of MCP (Model Context Protocol), it became possible to seamlessly integrate multiple live data sources like OP.GG and FPL into one intelligent system powered by Gemini.

What it does

Gemini with MCP Servers lets users ask natural language questions and get real-time esports and FPL data in a single chat interface. Whether it’s match schedules, player stats, or other insights, the system pulls information directly from multiple MCP servers and returns structured responses with the help of Gemini. The current focus is on esports (via OP.GG) and Fantasy Premier League, with future plans to extend to more sports.

How we built it

Google Gemini API Next.js and Tailwind CSS for the frontend. Node.js Server that connects to multiple MCP endpoints

Challenges we ran into

  1. We attempted to integrate Cloudbet for betting odds but ran into multiple schema errors in the MCP tool definitions, which was too time-consuming to fix during the build period.
  2. Coordinating multiple MCP transports while keeping the interface stable required careful mapping of tool names and arguments.

Accomplishments that we're proud of

  1. Successfully integrated two independent MCP servers (OP.GG and FPL).
  2. Built a fully working frontend that connects to Gemini and dynamically invokes tools exposed by MCP servers.
  3. Achieved smooth real-time streaming of responses using Streamable HTTP.

What we learned

  1. How to effectively use Streamable HTTP with MCP servers to enable low-latency, real-time responses.
  2. The importance of defining clear input schemas and arguments for tool calls, especially when connecting multiple MCP endpoints.

What's next for Gemini with MCP Servers

  1. Expanding to more sports, such as the NBA, NFL, and more FPL-like fantasy platforms.
  2. Re-integrating Cloudbet or similar services to support betting odds and predictions once schema issues are resolved.
  3. Adding richer UI components: team cards, visualizations, live match timelines.

Built With

Share this project:

Updates

Submission history