Skip to main content
Glama

MCPSmartProxy

Token-efficient MCP server gateway with semantic tool search

PyPI Python Coverage Ruff

mcp-name: io.github.daedalus/mcp-smart-proxy

Install

pip install mcp-smart-proxy

Related MCP server: ToolMux

Usage

from mcp_smart_proxy import ProxyConfig

config = ProxyConfig.from_file("proxy.yaml")

CLI

mcp-smart-proxy --help
mcp-smart-proxy serve --config proxy.yaml
mcp-smart-proxy serve --config proxy.yaml --watch ./servers/
mcp-smart-proxy serve --config proxy.yaml --transport streamable-http --host 0.0.0.0 --port 8000
mcp-smart-proxy index --config proxy.yaml
mcp-smart-proxy status --config proxy.yaml
mcp-smart-proxy validate --config proxy.yaml

The --watch option monitors a directory for .yaml, .yml, or .json files containing upstream server configurations. Add, modify, or remove files to dynamically update the available servers at runtime.

Transport Options

The --transport option supports:

  • stdio (default) - Standard input/output transport for local MCP clients

API

Config

  • ProxyConfig - Main configuration model

  • UpstreamConfig - Upstream server configuration

  • EmbeddingConfig - Embedding backend configuration

  • VectorStoreConfig - Vector store backend configuration

Models

  • ListResult - Result from the list tool

  • SearchResult - Result from the search tool

  • ToolResult - Result from tool calls

Development

git clone https://github.com/daedalus/mcp-smart-proxy.git
cd mcp-smart-proxy
pip install -e ".[test]"

# run tests
pytest

# format
ruff format src/ tests/

# lint
ruff check src/ tests/

# type check
mypy src/

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      D
      maintenance
      A universal gateway that aggregates multiple MCP servers into a single interface while providing advanced token optimization, result filtering, and automated summarization. It enables efficient management of large tool catalogs and reduces context usage by up to 95% for major AI clients.
      18
      MIT
    • A
      license
      B
      quality
      D
      maintenance
      Aggregates multiple MCP servers via a single interface with token optimization and multiple operating modes (Gateway, Meta, Proxy, Search, Code).
      7
      8
      MIT
    • A
      license
      A
      quality
      D
      maintenance
      Enables connecting multiple downstream MCP servers into a single gateway, with unified tool search, namespacing, and deferred loading to preserve context window.
      2
      6
      MIT