Introduction
MCP Agent Bridge exposes AI coding agents -- Claude Code, OpenAI Codex, and GitHub Copilot -- as Model Context Protocol (MCP) servers. This lets any MCP client call any agent through a standardized interface, enabling cross-agent code review, second opinions, and multi-model collaboration.
Why?
AI coding agents are powerful individually, but they each have blind spots. By bridging them behind MCP, you can:
- Get a second opinion -- Send your diff to a different model for review before merging
- Cross-validate -- Compare responses from Claude, Codex, and Copilot on the same question
- Structured reviews -- Get typed JSON output with verdict, issues by severity, and suggestions
- Automate workflows -- Build pipelines that route work to the best agent for each task
How It Works
┌──────────────┐ MCP (stdio) ┌─────────────────┐ CLI subprocess
│ MCP Client │ ◄──────────────────► │ MCP Server │ ──────────────────► claude -p
│ (any agent) │ or HTTP/SSE │ (bridge layer) │ codex exec
└──────────────┘ └─────────────────┘ copilot -p
Each bridge server:
- Registers MCP tools (
review,ask,code_review, etc.) - Receives tool calls from any MCP client
- Spawns the underlying CLI agent as a subprocess
- Parses and returns structured results
Architecture
In production, each server runs as a singleton background service behind an HTTP proxy (mcp-proxy). Multiple clients share a single instance of each agent -- no cold starts, no duplicated processes.
┌─────────────┐
│ Claude Code │──┐
└─────────────┘ │ HTTP/SSE ┌───────────┐ stdio ┌──────────────────┐
├──────────────────► │ mcp-proxy │ ──────────► │ claude-mcp-server│
┌─────────────┐ │ :8940/mcp └───────────┘ └──────────────────┘
│ Codex │──┘
└─────────────┘
┌─────────────┐
│ Claude Code │──────────────────► :8941/mcp ──► codex-mcp-server
└─────────────┘
┌─────────────┐
│ Claude Code │──────────────────► :8945/mcp ──► copilot-mcp-server
└─────────────┘
Key Features
| Feature | Description |
|---|---|
| Three agent bridges | Claude, Codex, and Copilot as MCP servers |
| Structured reviews | JSON output with verdict, issues, suggestions |
| Singleton services | Background daemons shared by multiple clients |
| Cross-platform | macOS (LaunchAgents), Windows (NSSM), Linux (systemd) |
| Standalone binaries | Zero-dependency executables compiled with Bun |
| Secure defaults | Read-only tools, localhost binding, input limits |
Next Steps
- Prerequisites -- What you need installed
- Installation -- Build and install the servers
- Quick Start -- Get running in 5 minutes