Custom MCP (Model Context Protocol) with ChatGPT AI Template

What is Custom MCP (Model Context Protocol) with ChatGPT AI Template?

Custom MCP (Model Context Protocol) with ChatGPT AI Template is an intelligent agentic system that integrates with Gmail and GitHub, using OpenAI models to autonomously decide which tools to use and how to process information. It is built for developers and users who need automated cross-service workflows.

How to use Custom MCP (Model Context Protocol) with ChatGPT AI Template?

After setting up the MCP server (FastAPI on port 8000) and the ChatGPT AI Template frontend (Next.js on port 3000), open the browser at http://localhost:3000, click "MCP Chat" in the sidebar, and enter a query. A provided start.sh script launches both servers together.

Key features of Custom MCP (Model Context Protocol) with ChatGPT AI Template

• Gmail integration: reads and analyzes emails from your Gmail account.
• GitHub integration: interacts with repositories to check issues and alerts.
• Intelligent decision making via OpenAI models.
• Autonomous operation between Gmail, GitHub, and OpenAI.
• Modern UI built with Next.js and Chakra UI.

Use cases of Custom MCP (Model Context Protocol) with ChatGPT AI Template

• Check if GitHub warnings exist in your Gmail inbox.
• Find dependency alerts for your repositories.
• Summarize recent GitHub notifications.
• Execute complex requests involving multiple services without manual steps.

FAQ from Custom MCP (Model Context Protocol) with ChatGPT AI Template

What external APIs does it require?

Gmail API (OAuth2), GitHub API (Personal Access Token), and OpenAI API.

How is the system started?

Run ./start.sh to launch both the MCP server and the ChatGPT AI Template frontend. Alternatively start each manually using the setup instructions.

Where are credentials stored?

In a .env file (not in the repository). A .env.example shows required variables.

What are the runtime requirements?

Python for the backend (FastAPI) and Node.js for the frontend (Next.js/React). Dependencies are listed in requirements.txt and package.json.