MCP server · Developer tools

ws-mcp (WebSocket MCP bridge)

Turn other MCP helper servers into a WebSocket connection your AI tools can talk to.

Developer tools For it and development Other Tools and Integrations

ws-mcp is a small bridge program. It takes other MCP helper servers, which normally talk over a plain text pipe on your computer, and puts them behind a WebSocket address instead. It is handy if you are using kibitz or another tool that expects a WebSocket connection rather than a local command.

What is an MCP server? The 30-second version

On its own, your AI can only chat. An MCP server is a small helper program that gives your AI a new skill, like searching the web or reading files. ws-mcp is a helper that takes other MCP helpers and re-publishes them over a WebSocket, which is a live two-way connection over the network. So instead of your AI launching each helper itself, it connects to ws-mcp at an address, and ws-mcp passes the requests through to the helpers you configured.

What this MCP server does

You start ws-mcp on your computer and tell it which MCP helper servers to run, either with a config file or by listing commands. ws-mcp launches those helpers and listens on a port, like 10125. When a tool such as kibitz connects to that port over WebSocket, ws-mcp forwards whatever the AI asks for to the right helper. The helper does the actual work, and the answer travels back through ws-mcp to your AI. You see the result in your chat as if the helper were connected directly.

Flow diagram: you ask your AI “Look up the weather and check my notes”, the ws-mcp (WebSocket MCP bridge) reaches Web pages, Files and system and Web search, and you get back answers in your chat.

Click to zoom

What you can do with it

Serve one or more MCP helper servers over a single WebSocket port
Pick which helpers to run using a config file or command flags
Pass environment variables and .env files through to the helpers
Run the fetch helper so your AI can make web requests
Run the wcgw helper for general system and file operations
Run the Brave search helper if you supply an API key
Choose a custom port for each WebSocket endpoint

Try asking your AI

“Start ws-mcp with the default config on port 10125”
“Run ws-mcp with the fetch helper on port 3002”
“Run ws-mcp with the Brave search helper using my BRAVE_API_KEY on port 3003”
“Run ws-mcp with two helpers at once using a .env file on port 3004”

What it gives back to you

ws-mcp itself does not produce answers. It passes your AI's requests to the helper servers you configured and sends their replies back over the WebSocket. What you see in the chat depends on the helper, for example a web page summary from fetch or a file listing from wcgw. If a helper is not running or the port is wrong, you will see a connection error instead.

Before you start

What you need

uv installed (a tool for running Python programs; the README shows a one-line install command)
A config file or one or more --command values naming the MCP helper servers you want to serve
An API key for helpers that need one, like BRAVE_API_KEY for Brave search
A client that speaks WebSocket MCP, such as kibitz

Good to know

The helpers you wrap can read files, run system commands, or use paid API keys, so only run ws-mcp with helpers you trust and keep the port off the public internet.

Install it with your AI

Add ws-mcp (WebSocket MCP bridge) to your AI, no technical skills needed

You don't install anything by hand. You copy one prompt, paste it into an AI that can work on your computer, and it checks, installs and connects the server for you, asking you when it needs something.

Sign in to get the install prompt

Members get a ready-made prompt that lets the Claude desktop app check ws-mcp (WebSocket MCP bridge), install it and connect it for them, step by step. You don't need any technical skills: you copy, paste and answer a few questions. Your connected AI can also find and install any of the 4,066 MCP servers here for you.