Claude Mcp Server Gateway

Name: Claude Mcp Server Gateway
Author: bzsasson

@bzsasson

About Claude Mcp Server Gateway

Discover the power of the “CLGateway”, an the open-source Claude MCP Server Gateway (available on GitHub). Built for users of Claude Desktop, Claude API or Claude Code, this Python-based gateway smartly manages your MCP (Model Context Protocol) servers and dramatically slashes to

#context-management #claude-desktop #token-optimization #claude-mcp #mcp-gateway #claude-mcp-server #mcp-wrapper #mcp-server-gateway

Basic information

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "claude-mcp-server-gateway": {
      "command": "/path/to/your/.venv/bin/python",
      "args": [
        "/path/to/claude-mcp-server-gateway/dcl_wrapper.py"
      ],
      "env": {}
    }
  }
}

Tools

No tools detected

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is Claude Mcp Server Gateway?

Claude Mcp Server Gateway is a Python MCP server gateway that loads MCP tools on demand instead of at startup, reducing token consumption. It integrates with Claude Desktop, Claude Code, Cline, and Roo, and works with any MCP server configured for those platforms.

How to use Claude Mcp Server Gateway?

Install via git, set up a Python 3.11+ virtual environment, install dependencies (mcp, python-dotenv), and configure API keys in a .env file. Then add the gateway as a single MCP server in your platform’s configuration file (e.g., claude_desktop_config.json, .claude.json, .mcp.json, or cline_mcp_settings.json) pointing to dcl_wrapper.py.

Key features of Claude Mcp Server Gateway

Loads MCP tools on demand instead of at startup
Exposes only 3 tools: list_available_mcps, load_mcp_tools, call_mcp_tool
Works with Claude Desktop, Claude Code, Cline, and Roo
Supports any MCP server configured in the gateway
Reduces token usage by deferring tool schema loading
Compatible with user-level and project-level configurations (Claude Code)

Use cases of Claude Mcp Server Gateway

Connect dozens of MCP servers without clogging context with tool definitions
Prevent MCP servers from starting until explicitly needed by Claude
Manage multiple MCP servers through a single gateway entry in configuration
Use with Claude Code CLI (claude mcp add) for rapid setup
Simplify onboarding of new MCP servers by editing only one Python dict

FAQ from Claude Mcp Server Gateway

What platforms does Claude Mcp Server Gateway support?

It supports Claude Desktop (app), Claude Code (terminal), Cline, and Roo (VSCode extensions). Each platform uses a specific configuration file location.

How does it reduce token usage?

Instead of loading all MCP tool definitions (100+ tools) on startup, the gateway exposes only 3 tools. Tools are loaded on demand when Claude calls load_mcp_tools for a specific server.

What are the runtime requirements?

Python 3.11+, the MCP Python SDK, and python-dotenv. The gateway uses a virtual environment and communicates over stdio with the host application.

What are the known limitations?

There is a small delay on the first tool use (server startup), each MCP server connection is temporary, and no persistence exists between calls.

How do I add a custom MCP server to the gateway?

Edit dcl_wrapper.py and add an entry to the MCP_SERVERS dictionary with the command, args, env, and description fields. The API key should be read from the .env file.

Comments

More AI & Agents MCP servers

MCP Client for Ollama (ollmcp)

jonigl

Harness the power of local LLMs with this TUI MCP Client for Ollama. Featuring all core MCP primitives (tools, prompts, resources), agent mode, multi-server, model switching, streaming responses, human-in-the-loop, thinking mode, model params config, system prompts, and saved pre

768

1Panel

1Panel-dev

🔥 1Panel is a modern, open-source VPS control panel — and the only one with native AI agent support. Run Ollama models, deploy OpenClaw agents, and manage your entire server stack from one clean web interface.

36.1K