# cob-rag-mcp

> cob-rag-mcp is an MCP server for knowledge and memory. MCP server and CLI that plug a self-hosted knowledge base into any agent. It runs locally on your machine and is started by your AI client.

- **Kind:** MCP server
- **Platforms:** CLI, self-hosted
- **Website:** https://callobuzz.github.io/cob-rag-mcp
- **Repository:** https://github.com/callobuzz/cob-rag-mcp
- **MCP Registry name:** io.github.callobuzz/cob-rag-mcp
- **Categories:** [knowledge and memory](https://ai-powered.com/categories/knowledge-and-memory), [AI models and LLM tooling](https://ai-powered.com/categories/ai-and-llm)
- **Listing updated:** 2026-11-18

## About

This package is purely a client for cob-rag-engine, a self-hosted retrieval layer you run yourself. It does not send data to a hosted service; the URL and key you provide decide which engine and which isolated knowledge base, called a corpus, the agent can reach. A corpus key (ck_) sees one knowledge base and nothing else, while a tenant key (tk_) manages more.

The agent calls search to pull passages with sources and answer with its own model, or ask for a complete cited answer written by the model configured in the engine. search needs no model at all. The same 52 tools also exist as CLI commands, such as search, ingest records, corpora create and usage, with tables by default and JSON for scripts.

Destructive tools are hidden unless COB_RAG_ALLOW_DESTRUCTIVE=true, and COB_RAG_READ_ONLY offers only read tools. It needs Node.js 22 or newer and speaks stdio and streamable HTTP.

## Features

- **52 shared tools:** Every MCP tool is also a CLI command.
- **Passage search:** Returns passages with sources and scores without needing a model.
- **Cited answers:** The ask tool returns an answer written by the engine's model.
- **Key-scoped access:** Corpus and tenant keys limit what a client can see.
- **Read-only mode:** COB_RAG_READ_ONLY exposes only reading tools.
- **Destructive tools hidden:** Destructive actions need an explicit environment variable.

## Good for

- Answer customer questions from a company's own policies
- Search internal documents from Claude or Cursor
- Ingest records into a knowledge base from a script
- Check engine usage from the shell

## Install

### Claude Code (Terminal)

```sh
claude mcp add cob-rag-mcp --env COB_RAG_URL='<COB_RAG_URL>' --env COB_RAG_API_KEY='<COB_RAG_API_KEY>' -- npx -y cob-rag-mcp
```

### Claude Desktop (claude_desktop_config.json)

```json
{
  "mcpServers": {
    "cob-rag-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "cob-rag-mcp"
      ],
      "env": {
        "COB_RAG_URL": "<COB_RAG_URL>",
        "COB_RAG_API_KEY": "<COB_RAG_API_KEY>"
      }
    }
  }
}
```

### Cursor (~/.cursor/mcp.json)

```json
{
  "mcpServers": {
    "cob-rag-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "cob-rag-mcp"
      ],
      "env": {
        "COB_RAG_URL": "<COB_RAG_URL>",
        "COB_RAG_API_KEY": "<COB_RAG_API_KEY>"
      }
    }
  }
}
```

### VS Code (.vscode/mcp.json)

```json
{
  "servers": {
    "cob-rag-mcp": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "cob-rag-mcp"
      ],
      "env": {
        "COB_RAG_URL": "<COB_RAG_URL>",
        "COB_RAG_API_KEY": "<COB_RAG_API_KEY>"
      }
    }
  }
}
```

### Windsurf (~/.codeium/windsurf/mcp_config.json)

```json
{
  "mcpServers": {
    "cob-rag-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "cob-rag-mcp"
      ],
      "env": {
        "COB_RAG_URL": "<COB_RAG_URL>",
        "COB_RAG_API_KEY": "<COB_RAG_API_KEY>"
      }
    }
  }
}
```

### Other clients (your client's MCP settings (most use this shape))

```json
{
  "mcpServers": {
    "cob-rag-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "cob-rag-mcp"
      ],
      "env": {
        "COB_RAG_URL": "<COB_RAG_URL>",
        "COB_RAG_API_KEY": "<COB_RAG_API_KEY>"
      }
    }
  }
}
```

## Tools

- `search`: Pull relevant passages with sources from a knowledge base.
- `ask`: Return a cited answer written by the engine's model.

## FAQ

### Is there a hosted service?

No. You host the engine yourself, for example with docker compose up.

### What must I configure?

COB_RAG_URL and COB_RAG_API_KEY.

### Which clients work?

Claude, Cursor, VS Code, n8n and any MCP client.

### What Node version is required?

Node.js 22 or newer.

### How do I add cob-rag-mcp to Claude Code?

Run this command in your terminal: claude mcp add cob-rag-mcp --env COB_RAG_URL='<COB_RAG_URL>' --env COB_RAG_API_KEY='<COB_RAG_API_KEY>' -- npx -y cob-rag-mcp

### What are the best alternatives to cob-rag-mcp?

The closest alternatives to cob-rag-mcp are Codebase Memory, XERJ, headroom, omniget and next-ai-draw-io.

## Where this comes from

- Description: written from the project's README and the project's website.
- Install steps: from the official MCP Registry entry.

## Alternatives

- [Codebase Memory](https://ai-powered.com/mcp/codebase-memory): Codebase Memory is an open-source MCP server for knowledge and memory.
- [XERJ](https://ai-powered.com/mcp/xerj): XERJ is an open-source MCP server for knowledge and memory.
- [headroom](https://ai-powered.com/mcp/headroom-3): headroom is an open-source MCP server for ai models and llm tooling.
- [omniget](https://ai-powered.com/mcp/omniget): omniget is an open-source MCP server for ai models and llm tooling.
- [next-ai-draw-io](https://ai-powered.com/mcp/next-ai-draw-io-2): next-ai-draw-io is an open-source MCP server for ai models and llm tooling.
- [mcp-for-blender](https://ai-powered.com/mcp/mcp-for-blender): mcp-for-blender is an open-source MCP server for design and media.
- [Claude Code Ultimate Guide](https://ai-powered.com/mcp/claude-code-ultimate-guide): Claude Code Ultimate Guide is an open-source MCP server for ai models and llm tooling.
- [unity-mcp](https://ai-powered.com/mcp/unity-mcp-2): unity-mcp is an open-source MCP server for ai models and llm tooling.

---
Source: https://ai-powered.com/mcp/cob-rag-mcp
