MCP Http Proxy

@spontaneous-order

Basic information

About MCP Http Proxy

An HTTP/SSE proxy server for Model Context Protocol (MCP) applications using stdio. Supports raw JSON-RPC commands via HTTP and implements direct stdio communication without an MCP SDK.

#mcp-http-proxy#http-proxy#json-rpc

MCP Http Proxy config

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

{
  "mcpServers": {
    "mcp-http-proxy": {
      "command": "node",
      "args": [
        "rpc-proxy-worker.js"
      ]
    }
  }
}

MCP Http Proxy tools

Tool list not available yet

This listing does not currently have a tool list to display. This does not mean the server has no tools; check its documentation for supported capabilities.

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

Overview

What is mcp-http-proxy?

mcp-http-proxy is an HTTP/SSE intermediary proxy server that manages a stdio-based MCP (Model Context Protocol) server process and exposes its tools and resources over standard HTTP and Server-Sent Events. It is designed for developers and applications that need to interact with MCP processes without handling stdio directly.

How to use mcp-http-proxy?

Run the script rpc-proxy-worker.js with Node.js (no additional install steps besides having Node.js). The server starts on port 3005 by default, serving a web dashboard at http://localhost:3005. Clients interact via HTTP endpoints (e.g., GET /tool/:toolName, POST /rpc/raw/command) or by connecting to the SSE endpoint at /sse.

Key features of mcp-http-proxy

Manages the lifecycle of a child MCP server process.
Provides HTTP endpoints for tools and raw JSON-RPC commands.
Delivers real-time events via Server-Sent Events (SSE).
Includes a web dashboard for tool discovery and debugging.
Translates URL query parameters into JSON-RPC tools/call requests.
Supports multiple client types (curl, browser, scripts).
Handles parameter validation and error relay.
Provides debugging endpoints for raw message history and logs.

Use cases of mcp-http-proxy

Expose a local stdio-only MCP server to web applications via HTTP.
Allow multiple HTTP clients to share a single managed MCP process.
Enable programmatic interaction with MCP tools using standard REST/SSE.
Debug and test MCP server responses without writing a custom client.

Frequently asked questions

What runtime does mcp-http-proxy require?

The proxy runs on Node.js. No other runtime dependencies are mentioned; the MCP server it manages is expected to be a separate process.

How do clients communicate with the MCP server through the proxy?

Clients send HTTP requests to endpoints like `/rpc/raw/command` (raw JSON-RPC) or `/tool/:toolName` (URL parameter style). The proxy forwards these as JSON-RPC commands to the child MCP process via stdin, and returns the response from stdout. Asynchronous events are pushed over SSE.

What endpoints does mcp-http-proxy provide?

The main endpoints are: `GET /` (dashboard), `GET /tools`, `GET /tool/:toolName`, `POST /rpc/raw/command`, `GET /sse`, `GET /help`, and several debugging endpoints (`/rpc/command`, `/rpc/raw`, `/debug`).

How does the proxy handle errors?

It returns HTTP status codes (e.g., 404 for unknown tools, 400 for invalid parameters) and validates parameters for `/tool/:toolName`. Errors from the MCP process are relayed back in HTTP responses or pushed via SSE.

Does mcp-http-proxy support authentication?

The README does not mention any authentication mechanism. The proxy is designed for local or trusted network use.

Comments

More Developer Tools MCP servers

Moxie Docs

Jackalope-Dev

Automated codebase documentation for GitHub: searchable docs, MCP context for AI agents, doc drift detection, and Friday Cleanup PRs.

The Bridge

Complete, hosted login, teams, billing and feature flags for your SaaS app, set up by your AI assistant: sign-in and SSO, customer teams and roles, plans and usage limits on Stripe, and feature flags. OAuth sign-in, no A

LastPing

Know when your AI agents, cron jobs and CI pipelines fail, stall or go quiet. Your agent sets up monitors from a plain description, tells you what is broken and why, reads run history and traces, and writes notes on inci

Mcp-Warden

ermoluk

Local-first credential broker for MCP. AI coding agents (Claude Code, Cursor, Codex…) call APIs, SSH and databases through scoped presets; secrets stay in your OS keychain and never enter the agent's context.

Modelglass

Modelglass turns "which model should I use?" into a lookup instead of a guess. Seven tools give your agent live pricing across image, language, video, and audio models. Sourced and versioned, not scraped once and forgott