proxy-mcp

Universal MCP proxy server that discovers, searches, and executes tools across all your configured MCP servers from a single entry point.

Features

• Proxy execution — Call any tool on any MCP server through execute_mcp_tool
• Semantic search — Find tools by describing what you want to do, not by exact name
• Connection pool — MCP subprocesses are spawned on-demand and reused across calls
• Weight learning — The search system learns from your selections to improve rankings over time
• Auto-indexing — All MCP tools are indexed on server startup (non-blocking)

Tools

Tool	Description
execute_mcp_tool	Execute any tool on any configured MCP server
search_tools	Semantic search across all indexed tools
list_available_mcps	List configured MCP servers and their status
refresh_tool_index	Re-index tools from MCP servers
get_learned_weights	View learned search weights
reset_learned_weights	Reset all learned weights
close_mcp_connections	Close all active MCP connections

Requirements

• Python >= 3.10
• uv (recommended) or pip

Installation

# Clone
git clone https://github.com/Rixmerz/proxy-mcp.git
cd proxy-mcp

# Install with uv
uv sync

Configuration

Add to ~/.claude.json under mcpServers:

{
  "mcpServers": {
    "proxy-mcp": {
      "command": "uv",
      "args": ["run", "--directory", "/path/to/proxy-mcp", "proxy-mcp"]
    }
  }
}

The proxy reads MCP server configs from:

1. ~/.agentcockpit/mcps.json (priority)
2. ~/.claude.json (fallback)

Usage examples

# Search for a tool by what you want to do
search_tools(query="expose service to internet")

# Execute a tool on a specific MCP
execute_mcp_tool(
    mcp_name="Context7",
    tool_name="get-library-docs",
    arguments={"context7CompatibleLibraryID": "/vercel/next.js", "topic": "routing"}
)

# List all available MCPs
list_available_mcps()

Data

Learned weights are stored at ~/.proxy-mcp/learned_weights.json.

License

MIT