MCP-AI-App-with-Gemini

wuwen1030/MCP-AI-App-with-Gemini

3.2

If you are the rightful owner of MCP-AI-App-with-Gemini and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.

This document provides a structured summary of a Model Context Protocol (MCP) server designed for integrating AI applications with Google Gemini and custom research functionalities.

Tools
2
Resources
0
Prompts
0

MCP: Build AI App with Gemini

A demonstration project for building MCP-enabled chatbots with Google Gemini integration and a custom research MCP server.

Inspired by MCP: Build Rich-Context AI Apps with Anthropic.
This project provides an alternative for developers who cannot access Anthropic's Claude API in certain regions.

Overview

This project provides a complete demonstration of how to build both MCP (Model Context Protocol) clients and servers. It showcases the full MCP ecosystem by implementing:

MCP Client Side:

  • Chatbot powered by Google Gemini for intelligent conversations.
  • Implements MCP core concepts on the client: tools, resources, and prompt templates.
  • Connects to multiple MCP servers via a simple configuration file.

MCP Server Side:

  • A custom research server for arXiv paper search and management
    • Tools: search_papers(topic, max_results) and extract_info(paper_id)
    • Resources: Browse saved papers via papers://folders and papers://{topic}
    • Prompt Templates: generate_search_prompt(topic, num_papers) for structured literature review

Installation

Prerequisites

  • Python 3.13+
  • Node.js (for filesystem MCP server)
  • uv (Python package manager)

Setup

  1. Clone the repository:

  2. Install Python dependencies:

    uv sync

  3. Set up environment variables:

    cp env.example .env
# Edit .env and add your GEMINI_API_KEY

  4. Get your Gemini API key:

    • Visit Google AI Studio
    • Create a new API key
    • Add it to your .env file: 
      GEMINI_API_KEY=your_api_key_here

Usage

Running the Chatbot

Start the interactive chatbot:

uv run client.py

The chatbot will:

  1. Connect to all configured MCP servers
  2. Initialize the Gemini AI model
  3. Start an interactive chat session

Interactive Commands

Inside the chat session, you can use:

  • @folders: List available topics that have saved papers (papers://folders).
  • @: Show a markdown summary of papers for a topic (papers://{topic}). Example: @agentic_workflows.
  • /prompts: List all available prompt templates.
  • /prompt <arg=value> ...: Execute a prompt and let Gemini continue the conversation.
    • Example: /prompt generate_search_prompt topic="graph neural networks" num_papers=5

You can also ask questions in natural language. Gemini will automatically call tools the servers expose when appropriate.

What’s Included

  • Research MCP server (research_server.py)
    • Tools
      • search_papers(topic, max_results=5): Fetches papers from arXiv and stores metadata locally under papers/<topic>/papers_info.json.
      • extract_info(paper_id): Looks up a paper across all topics and returns its saved metadata.
    • Resources
      • papers://folders: Lists all topics with saved papers.
      • papers://{topic}: Renders a markdown overview of papers saved for that topic.
    • Prompt Templates
      • generate_search_prompt(topic, num_papers=5): Produces a structured research prompt that instructs the model to search, extract, and synthesize findings.

Multiple Servers

By default, server_config.json connects to:

  • filesystem: Node-based MCP server for local file browsing (requires Node.js). It may not implement all MCP methods (e.g., tools list), which is normal.
  • research: Python MCP server in this repo for arXiv research workflows.
  • fetch: Generic HTTP fetch MCP server.

The client handles optional capabilities gracefully, so you might see notices like [filesystem] tools unsupported: Method not found when a server doesn’t support a specific method.

Acknowledgments