MCP-OS · Model Context Protocol Orchestration System
About
MCP‑OS fetches just the MCPs your task needs, cutting prompt bloat and toggling servers on‑demand for a lean, secure toolset.
Details
- Author
- giao-123-sun
- GitHub stars
- 3
- Downloads
- 449
- Categories
- Productivity
Jump to
- Vector retrieval of top‑k MCPs from a local index.
- Slim prompt template that reduces prompt tokens by ~70%.
- Pluggable vector store backends (FAISS, Qdrant, Milvus, etc.).
- Default embedding backend using OpenAI embeddings.
- REST endpoint (/match) for task‑to‑MCP matching.
- Roadmap includes health‑check daemon, runtime manager, and policy sandbox.
Clone the repository, install dependencies (npm install), build the vector index from an MCP list (npm run build:index --src ./mcp_list.json --out ./index), then start the retriever server (npm run start:retriever), which listens on 127.0.0.1:5500 over HTTP+SSE. Wire it into your LLM/agent by configuring mcpServers (e.g., in Claude Desktop) or by calling the REST endpoint POST /match with a task description.
MCP-OS · Model Context Protocol Orchestration System
> Let your large language model focus on solving tasks—not wading through a sea of MCPs.> 👉 查看中文版 README
---
✨ Project Vision
As the Model Context Protocol (MCP) ecosystem explodes, hundreds of MCP servers create three familiar headaches:| Pain Point | Description |
| ---------- | ----------- |
| Prompt Bloat | Lengthy MCP descriptions crowd the context window; the model spends more tokens picking tools than planning / analysis. |
| Connection Hygiene | We must constantly track which MCPs are alive and whether they satisfy the current task. |
| Resource & Security | Always-on MCP servers consume memory and expose interfaces, increasing attack surface. |
MCP-OS aims to:
> “Manage MCPs the way an operating system manages processes—load on demand, unload when idle.”
---
🌟 Current Phase: MCP-Retriever (Completed ✅)
1. Vector Retrieval — Embed task descriptions and retrieve Top-k MCPs from a vector index. 2. Slim Prompt Template — Inject only the Top-k MCP descriptions, saving ~70 % prompt tokens on average. 3. Pluggable Back-ends — Defaultopenai/embeddings; swap in FAISS, Qdrant, Milvus, etc.
> 📖 Details in /packages/retriever.
---
🛣️ Roadmap
| Milestone | Feature | Status |
| --------- | ------- | ------ |
| v0.1 | MCP-Retriever – vector search | ✅ Released |
| v0.2 | MCP-Retriever - light version | ⏳ In progress |
| v0.3 | Health-Check Daemon – auto heartbeat & pruning | ⏳ In progress |
| v0.4 | Runtime Manager – on-demand MCP start/stop | 🗓 Planned |
| v1.0 | Policy Sandbox – fine-grained auth, rate, cost | 🗓 Planned |
---
⚙️ Quick Start
1. Clone & Install
git clone https://github.com/your-org/mcp-os.git
cd mcp-os
npm install # or npm / yarn
2. Build the Vector Index
```bash
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
