# GitHub MCP Server

GitHub MCP Server is a Model Context Protocol (MCP) server that connects AI tools directly to GitHub's platform. It is open source and self-hosted, installable on Docker Hub and Kubernetes (Helm).

Canonical page: https://ourlocalhost.com/app/5130/github-mcp-server

## Description

GitHub MCP Server is a Model Context Protocol (MCP) server that connects AI tools directly to GitHub's platform. It lets AI agents, assistants, and chatbots read repositories and code files, manage issues and pull requests, analyze code, and automate workflows through natural language. It is for developers and teams using AI coding assistants such as VS Code, Claude, Cursor, or Windsurf who want to manage GitHub work without leaving their editor.

## Platform availability

Available on 1 of 9 platforms: Helm.

## Health score

79/100 (Good), based on 3 of 4 factors.
- Maintained: 100 (last commit 25 days ago)
- Popular: 90 (33,271 GitHub stars)
- Easy to install: 25 (1 install platform, install notes)

## Links

- GitHub: https://github.com/github/github-mcp-server
- Docker Hub: https://hub.docker.com/u/mcp
- Kubernetes (Helm): https://artifacthub.io/packages/helm/mcp-charts/mcp

## Common questions

### How do I choose the listening port or host for an MCP server, and when is a tunnel appropriate?

Use mcp-use start --port 8080 or --host 0.0.0.0 to change where the server listens. Add --tunnel only if you intentionally need a public link, because it exposes the server.

### How can I troubleshoot an MCP server that fails during startup?

Run mcp-use build first, then mcp-use start --with-inspector. Building prepares the server files, and the inspector helps spot startup problems.

### How can I manage many MCP servers without duplicating client configuration?

MCPJungle can run multiple MCP servers behind one endpoint as a self-hosted MCP gateway. This reduces scattered client configurations, duplicated setup, and inconsistent access control.

### What should I watch for when an MCP server acts on behalf of users?

The most common and severe risk is the Confused Deputy Problem. This occurs when the MCP server acts with its own high-level permissions on behalf of a lower-privileged user.

## Alternatives

- Browserbase: Browserbase is a platform that lets AI tools surf the internet just like a human.
- puppeteer: Puppeteer is a Node library which provides a high-level API to control headless Chrome or Chromium over the DevTools Protocol.

---
Source: Our Local Host (https://ourlocalhost.com/app/5130/github-mcp-server), the self-hosted and open-source app directory.