agent and mcp-proxy
These are **competitors** — both aggregate multiple MCP servers into a single unified interface, with the primary difference being that mcp-app/agent targets direct MCP server integration while mcp-proxy/mcp-proxy adds an HTTP abstraction layer, but they solve the same core problem of MCP server consolidation and you would choose one based on your deployment architecture rather than use both together.
About agent
1mcp-app/agent
A unified Model Context Protocol server implementation that aggregates multiple MCP servers into one.
Implements stdio transport isolation with subprocess management for child MCP servers, featuring OAuth 2.1 authentication, hot-reload configuration, and health check monitoring. Designed as a proxy layer that aggregates tools across multiple AI assistants (Claude Desktop, Cursor, VSCode) through a single HTTP endpoint, eliminating per-client configuration overhead.
About mcp-proxy
tbxark/mcp-proxy
An MCP proxy server that aggregates and serves multiple MCP resource servers through a single HTTP server.
Supports multiple MCP client transport types (stdio, SSE, streamable-http) and exposes aggregated tools, prompts, and resources via Server-Sent Events or streamable HTTP endpoints. Built in Go with Docker support for launching Node.js and Python-based MCP servers via `npx` and `uvx`. Includes a web-based Claude configuration converter for simplified client setup.
Related comparisons
Scores updated daily from GitHub, PyPI, and npm data. How scores work