AI Agent with Model Context Protocol (MCP) Server of HTTP with SSE and Google Gemini 2.5 Flash Model

Overview

This project implements an AI Agent server that utilizes the Model Context Protocol (MCP) over HTTP, enhanced with Server-Sent Events (SSE) for real-time communication. It is designed to facilitate seamless and efficient client-server interactions for AI-based tasks, leveraging the power of the Google Gemini 2.5 Flash Model for advanced natural language processing.

The server is built to act as a bridge between AI model inference and client-side applications, providing streaming responses, low-latency connectivity, and scalable integration for a variety of use cases, such as chatbots, code review agents, and other intelligent assistants.

Key Features

• Model Context Protocol (MCP) Implementation:
  The server follows the MCP standard, allowing structured communication between clients and the AI agent, including context management, prompt exchange, and response tracking.

• HTTP & Server-Sent Events (SSE):
  Real-time, one-way communication from server to client using SSE, providing live streaming of AI responses and status updates.

• Google Gemini 2.5 Flash Model Integration:
  Out-of-the-box support for invoking the Gemini 2.5 Flash model, enabling state-of-the-art language and code understanding capabilities.

• Extensible & Modular:
  Designed for easy integration with other models, protocols, and front-end clients.

• Secure & Scalable:
  Leverages HTTP standards and can be deployed in modern cloud environments with scalability and security best practices.

Architecture

graph TD
    Client[Client] -- HTTP/SSE --> Server[MCP HTTP with SSE Server]
    Server -- API Call --> Gemini[Google Gemini 2.5 Flash Model]

MCP of HTTP with SSE Server's Workflow

• Client: Sends prompts/requests and receives streaming responses via SSE.
• MCP SSE Server: Handles protocol logic, manages context, and communicates with the Gemini model.
• Google Gemini Model: Provides AI inference and generates responses.

Use Cases

• AI-powered chatbots with real-time streaming
• Automated regular basic tasks.

Getting Started

Prerequisites

• Node.js (or language/runtime as per your implementation)
• Access to Google Gemini 2.5 Flash API (API key/configuration)
• Git

Installation

git clone https://github.com/sabbirkhanoni/AI-Agent-with-Model-Context-Protocol-MCP-Server-of-HTTP-with-SSE-and-Google-Gemini-2.5-Flash-Model.git
cd AI-Agent-with-Model-Context-Protocol-MCP-Server-of-HTTP-with-SSE-and-Google-Gemini-2.5-Flash-Model
npm install

Configuration

Create a .env file and set your Gemini API key and other environment variables as needed:

GEMINI_API_KEY=your_google_gemini_api_key
PORT=****

Running the Server

npm start

The server will be available at http://localhost:****.

Contributing

Contributions, issues, and feature requests are welcome!
Please open an issue or submit a pull request.

Acknowledgements

• Google Gemini 2.5 Flash Model
• Model Context Protocol (MCP)