broxy
Broxy fronts multiple upstream MCP servers as one, aggregates capabilities across protocols, and exposes only what presets allow.
Install / Use
claude mcp add qent -- npx -y github:qent/broxyIf the server publishes to npm under a different name, use that package instead — check the repo README.
MCP Server
Model Context Protocol server
Quality Score
Category
Development & EngineeringSupported Platforms
Tags
Skill content
View source on GitHubWhy Broxy
<ul style="list-style: none; padding-left: 0;"> <li>💰 Save tokens and money by exposing only the essential tools, prompts, and resources per task.</li> <li>🎯 Build more efficient agents with focused toolsets and fewer distractions.</li> <li>🔌 Connect ChatGPT and other HTTP-only MCP clients to local STDIO MCP servers.</li> <li>⚡ Instantly context-switch by hot-swapping presets.</li> <li>🔐 Sign in with Google to access the dashboard and remote connectivity features.</li> <li>📄 Apache 2.0 licensed for personal or commercial use.</li> </ul>Features
<table role="presentation"> <tr> <td> <img src="https://broxy.run/static/resources/features/mcp_list.png" alt="Set up your MCP servers once, use it with any client" /> <h3>Set up your MCP servers once</h3> Centralize MCP servers without duplicating per-client settings. Connect local or remote servers over STDIO, HTTP, SSE, or WebSocket, and add custom icons to keep large lists readable. <br /> </td> <td> <img src="https://broxy.run/static/resources/features/preset_creation.png" alt="Mix capabilities from servers into one preset" /> <h3>Mix capabilities from MCP servers</h3> Build purpose-built presets by mixing tools, prompts, and resources from any connected server. Each preset exposes only what an agent needs and reduces context noise. <br /> </td> </tr> <tr> <td> <img src="https://broxy.run/static/resources/features/chatgpt_to_local_mcp.png" alt="Connect ChatGPT to local MCP servers" /> <h3>Connect ChatGPT to local MCP servers</h3> Use local MCP servers inside ChatGPT on web, desktop, or mobile. Broxy Cloud bridges the connection securely so private tools are available anywhere you sign in. <br /> </td> <td> <img src="https://broxy.run/static/resources/features/popular_clients.png" alt="Connect to popular clients in one click" /> <h3>Connect to popular clients in one click</h3> Auto-configure popular AI clients across HTTP, SSE, and STDIO modes. Broxy updates only the MCP section so the rest of your client configuration stays untouched. <br /> </td> </tr> <tr> <td> <img src="https://broxy.run/static/resources/features/dynamic_client_with_oauth.png" alt="Dynamic client registration with OAuth" /> <h3>Dynamic client registration with OAuth</h3> Authorize servers that require personal data without manual client setup. Broxy discovers OAuth metadata, registers clients dynamically, and works even when the target client lacks OAuth support. <br /> </td> <td> <img src="https://broxy.run/static/resources/features/adapter_mode.png" alt="Stable MCP surface for preset changes" /> <h3>Stable MCP surface for preset changes</h3> Expose a stable MCP surface so clients do not need to refresh capabilities when presets change. Adapter mode also surfaces prompts and resources for clients that only support tools, including ChatGPT. <br /> </td> </tr> </table>Connections
STDIO (local connection)
Use STDIO when your client expects a local MCP process.
{
"mcpServers": {
"broxy": {
"command": "/Applications/broxy.app/Contents/MacOS/broxy",
"args": [
"--stdio-proxy"
]
}
}
}
Streamable HTTP
Broxy exposes Streamable HTTP on http://localhost:3335/mcp by default.
{
"mcpServers": {
"broxy": {
"url": "http://localhost:3335/mcp"
}
}
}
SSE inbound
SSE is served on the same host and port at /sse.
GET http://localhost:3335/sseopens the SSE stream and returns the session endpoint.POST http://localhost:3335/sse?sessionId=...forwards MCP JSON-RPC messages into the session.
Developer Documentation
For contributors and AI code agents:
- Start with
docs/readme.md(2-level docs index). - Use
docs/agent_quickstart.mdfor task-to-doc mapping, invariants, and required checks. - Canonical runtime contracts are in:
docs/core_contracts.mddocs/proxy_facade.mddocs/inbound_transports.mddocs/presets_and_filtering.mddocs/configuration_and_hot_reload.md
CLI jar
Build the jar
./gradlew :cli:shadowJar
The jar is generated at cli/build/libs/broxy-cli.jar.
Run STDIO
java -jar cli/build/libs/broxy-cli.jar proxy \
--preset-id <preset-id> \
--config-dir ~/.config/broxy \
--inbound stdio
Run Streamable HTTP + SSE
java -jar cli/build/libs/broxy-cli.jar proxy \
--preset-id <preset-id> \
--config-dir ~/.config/broxy \
--inbound http \
--url http://localhost:3335/mcp
The CLI expects mcp.json and preset_<id>.json in the config directory (default ~/.config/broxy).
Related Skills
Agent-Reach
77.4kGive your AI agent eyes to see the entire internet. Read & search Twitter, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu — one CLI, zero API fees.
ruflo
70.1k🌊 The original agent meta-harness. Deploy intelligent multi-player swarms, coordinate autonomous workflows, and build conversational AI systems. Features adaptive memory, self-learning intelligence, RAG integration, and native Claude Code / Codex / Hermes and many more Integrated
headroom
68.3kCompress tool outputs, logs, files, and RAG chunks before they reach the LLM. 20% fewer tokens for coding agents, 60-95% fewer tokens for JSON, same answers. Library, proxy, MCP server.
CowAgent
46.8kOpen-source super AI assistant & Agent Harness. Plans tasks, runs tools and skills, self-evolves with memory and knowledge. Multi-model, multi-channel. Lightweight, extensible, one-line install. (formerly chatgpt-on-wechat)
