mcphub  by samanhappy

Unified hub server for aggregating multiple MCP servers

created 4 months ago
975 stars

Top 38.7% on sourcepulse

GitHubView on GitHub
Project Summary

MCPHub is a unified server hub designed to aggregate multiple Model Context Protocol (MCP) servers, exposing them via separate Streamable HTTP (SSE) endpoints. It targets AI developers and researchers seeking to streamline integrations with various AI tools and services through a centralized, configurable platform.

How It Works

MCPHub acts as a proxy and management layer for diverse MCP-compatible backend services (e.g., amap-maps, playwright, fetch, slack). It leverages Node.js and Express for its backend, with a React frontend for a dashboard. The core functionality revolves around its flexible protocol handling, supporting both standard I/O and SSE MCP protocols, and offering hot-swappable configuration for seamless server management without downtime.

Quick Start & Requirements

  • Install/Run: docker run -p 3000:3000 -v $(pwd)/mcp_settings.json:/app/mcp_settings.json samanhappy/mcphub (recommended) or docker run -p 3000:3000 samanhappy/mcphub for default settings.
  • Prerequisites: Docker. For local development: Node.js, pnpm.
  • Configuration: Optional mcp_settings.json for server definitions and user management. Default credentials: admin/admin123.
  • Docs: Dashboard Overview

Highlighted Details

  • Unified HTTP endpoint (/mcp) and group-specific endpoints (/mcp/{group}) for simplified AI client integration.
  • Centralized dashboard for real-time monitoring, server management, and user administration.
  • Hot-swappable configuration allows adding/removing MCP servers without service interruption.
  • Built-in JWT and bcrypt for secure user authentication and role-based access control.

Maintenance & Community

  • Active development with contributions welcome.
  • Discord community available for discussions and support.
  • Star History available.

Licensing & Compatibility

  • Licensed under the Apache 2.0 License.
  • Permissive license suitable for commercial use and integration with closed-source applications.

Limitations & Caveats

The README notes that support for streaming HTTP endpoints varies across AI clients, recommending the SSE endpoint as a fallback. The SSE endpoint is also marked as deprecated in future updates, indicating a potential shift towards HTTP-only streaming.

Health Check
Last commit

4 days ago

Responsiveness

Inactive

Pull Requests (30d)
17
Issues (30d)
18
Star History
864 stars in the last 90 days

Explore Similar Projects

Feedback? Help us improve.