mcp-reticle

Reticle is a proxy + UI that lets you observe, correlate, and profile MCP traffic in real time. It can wrap an existing MCP server that communicates over stdio, then expose a live view of requests, notifications, and responses, along with latency and token estimates. You can also run Reticle in log-only mode or proxy HTTP-based MCP servers to inspect traffic via the UI. Use the wrap command to start Reticle around your MCP server so you can see everything as it flows between your client and server. The included UI supports visualizing the transport layers (stdio, HTTP/SSE, WebSocket, etc.) and exporting captured sessions for later analysis. The CLI also offers a headless telemetry hub for collecting and streaming metrics.

Prerequisites:

• Node.js and npm (recommended): install from https://nodejs.org
• Optional: Python and pip if you plan to use the Python client or examples
• Optional: Homebrew for macOS users

Installation options:

1. Install the MCP proxy UI globally via npm: npm install -g mcp-reticle

2. Install via pip (Python): pip install mcp-reticle

3. Install via Homebrew (macOS): brew install labterminal/mcp-reticle/mcp-reticle

4. Build from source (advanced): git clone https://github.com/labterminal/mcp-reticle.git cd mcp-reticle just build

Basic usage after installation:

• Wrap an MCP server: mcp-reticle run --name my-server -- <your-server-command>
• Launch the UI: mcp-reticle ui
• Run in log-only mode (no UI): mcp-reticle run --log -- <your-server-command>
• Proxy an HTTP-based MCP server: mcp-reticle proxy --name api --upstream http://localhost:8080 --listen 3001

Tips and notes:

• Reticle supports multiple transports: stdio, Streamable HTTP, WebSocket, and HTTP/SSE. Ensure your MCP server is reachable by Reticle through the chosen transport.
• When wrapping a server, you can give it a descriptive name with --name for easier identification in the UI.
• The --log flag enables a lightweight, non-UI mode suitable for headless logging environments.
• For large sessions, export logs from the UI or the CLI to share with team members.
• If you encounter transport-specific issues, consult the CLI reference and wiki for transport configuration tips and troubleshooting steps.
• The current project ships as both an npm and a Python package; choose the installation method that aligns with your environment and workflow.