Skip to main content
Glama

Code Mode Bridge

npm

An MCP (Model Context Protocol) multiplexer that connects to upstream MCP servers and exposes all their tools through a single sandbox_eval_js tool for unified orchestration and execution. Runs on Node.js and Deno. Bun is not officially supported — if you need to use Bun, ensure a container runtime (Docker or Podman) or Deno is available in your PATH so the bridge can delegate code execution to a supported executor.

Key Features

  • Multi-server bridging: Connect to multiple upstream MCP servers simultaneously.

  • Tool aggregation: Exposes all upstream functions through a single sandbox_eval_js tool.

  • Dynamic discovery: Three built-in discovery tools let agents find and inspect available functions before writing code.

  • Multi-executor sandbox: Secure execution via Deno, isolated-vm, Docker/Podman containers, or vm2.

  • Automatic detection: Automatically selects the best available executor for your environment.

  • Live reload: Watches mcp.json for changes and hot-reloads upstream connections without restarting.

  • CLI management: Easy command-line interface for server configuration.

Related MCP server: Useful Helpers Suite

Quick Start

Installation & Run

# Run without installing
npx @ruifung/codemode-bridge

# Or install globally
npm install -g @ruifung/codemode-bridge
codemode-bridge run

Basic Usage

  1. Add a server:

    codemode-bridge config add kubernetes --type stdio --command "npx" --args "-y,kubernetes-mcp-server@latest"
  2. Start the bridge:

    codemode-bridge run
  3. Use in your client: Point your MCP client (Claude Desktop, VS Code, etc.) to npx @ruifung/codemode-bridge.

Discovery Tools

The bridge exposes three tools for discovering what's available before writing sandbox_eval_js code:

Tool

Description

sandbox_get_functions

List all available functions grouped by server. Accepts an optional server filter.

sandbox_get_function_schema

Get the TypeScript type definition for a specific function by name.

sandbox_search_functions

Keyword-search all function names and descriptions. Returns matching functions with their schemas.

Use these before calling sandbox_eval_js to find the correct function name and parameter types.

Using the sandbox_eval_js Tool

The sandbox_eval_js tool executes JavaScript in a sandboxed environment with access to all upstream functions via the codemode object. You can write either a function body or a complete async arrow function:

// Function body (auto-wrapped):
const result = await codemode.server_name__function_name({ param: "value" });
return result;

// Or as a complete async arrow function:
async () => {
  const result = await codemode.server_name__function_name({ param: "value" });
  return result;
}

Documentation Index

  • Architecture: How the bridge works, project structure, and logical flow.

  • Executors: Detailed comparison of sandbox backends (Deno, isolated-vm, Containers).

  • CLI & Configuration: Full command reference and configuration format.

  • HTTP Serve Mode: Multi-client HTTP transport, session isolation, and security.

  • Integration Guide: Setup instructions for Claude Desktop, VS Code, and more.

  • Development: Build and test instructions for contributors.

  • Sandbox Hardening: Security details for the execution environments.

A
license - permissive license
-
quality - not tested
C
maintenance

Maintenance

Maintainers
Response time
0dRelease cycle
12Releases (12mo)
Commit activity

Resources

Unclaimed servers have limited discoverability.

Looking for Admin?

If you are the server author, to access and configure the admin panel.

Related MCP Servers

  • F
    license
    -
    quality
    B
    maintenance
    A centralized MCP server providing over 40 tools for contextual intelligence, code analysis, vision, and memory management via MemPalace, enabling efficient project development and agent orchestration.
  • A
    license
    A
    quality
    C
    maintenance
    Aggregates multiple MCP servers into a single interface, reducing token overhead and simplifying tool management for LLMs.
    6
    54
    16
    Apache 2.0

View all related MCP servers

Related MCP Connectors

  • MCP server exposing the Backtest360 engine API as tools for AI agents.

  • A paid remote MCP for OpenAI Codex agent coordination MCP, built to return verdicts, receipts, usage

  • MCP server for AI agents to plan, verify, and deploy Cloudflare-native apps.

View all MCP Connectors

Latest Blog Posts

MCP directory API

We provide all the information about MCP servers via our MCP API.

curl -X GET 'https://glama.ai/api/mcp/v1/servers/ruifung/mcp-cm-bridge'

If you have feedback or need assistance with the MCP directory API, please join our Discord server