# Darwin RAG MCP server

> Enables AI agents to search and ingest local documents like PDFs, notes, and spreadsheets using a private, local-first RAG engine.

- Repository: https://github.com/Brightdotdev/DARWIN
- Page: https://hostedmcp.org/servers/brightdotdev/darwin
- Transport: stdio
- Stars: 3
- License: MIT
- Latest version: 0.2.18

## Install

### Claude Code

```sh
claude mcp add --env ANTHROPIC_API_KEY=<YOUR_ANTHROPIC_API_KEY> --env GEMINI_API_KEY=<YOUR_GEMINI_API_KEY> --env OPENAI_API_KEY=<YOUR_OPENAI_API_KEY> darwin -- uvx darwin-rag
```

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

```json
{
  "mcpServers": {
    "darwin": {
      "command": "uvx",
      "args": [
        "darwin-rag"
      ],
      "env": {
        "ANTHROPIC_API_KEY": "<YOUR_ANTHROPIC_API_KEY>",
        "GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
        "OPENAI_API_KEY": "<YOUR_OPENAI_API_KEY>"
      }
    }
  }
}
```

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

```json
{
  "servers": {
    "darwin": {
      "type": "stdio",
      "command": "uvx",
      "args": [
        "darwin-rag"
      ],
      "env": {
        "ANTHROPIC_API_KEY": "<YOUR_ANTHROPIC_API_KEY>",
        "GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
        "OPENAI_API_KEY": "<YOUR_OPENAI_API_KEY>"
      }
    }
  }
}
```

### Claude Desktop (`claude_desktop_config.json`)

```json
{
  "mcpServers": {
    "darwin": {
      "command": "uvx",
      "args": [
        "darwin-rag"
      ],
      "env": {
        "ANTHROPIC_API_KEY": "<YOUR_ANTHROPIC_API_KEY>",
        "GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
        "OPENAI_API_KEY": "<YOUR_OPENAI_API_KEY>"
      }
    }
  }
}
```

### Codex (`~/.codex/config.toml`)

```toml
[mcp_servers.darwin]
command = "uvx"
args = ["darwin-rag"]

[mcp_servers.darwin.env]
ANTHROPIC_API_KEY = "<YOUR_ANTHROPIC_API_KEY>"
GEMINI_API_KEY = "<YOUR_GEMINI_API_KEY>"
OPENAI_API_KEY = "<YOUR_OPENAI_API_KEY>"
```

### Gemini CLI (`~/.gemini/settings.json`)

```json
{
  "mcpServers": {
    "darwin": {
      "command": "uvx",
      "args": [
        "darwin-rag"
      ],
      "env": {
        "ANTHROPIC_API_KEY": "<YOUR_ANTHROPIC_API_KEY>",
        "GEMINI_API_KEY": "<YOUR_GEMINI_API_KEY>",
        "OPENAI_API_KEY": "<YOUR_OPENAI_API_KEY>"
      }
    }
  }
}
```

## Environment variables

| Name | Required | Secret | Description |
|---|---|---|---|
| `ANTHROPIC_API_KEY` | no | yes | Anthropic API key for LLM answer generation |
| `GEMINI_API_KEY` | no | yes | Google Gemini API key |
| `OPENAI_API_KEY` | no | yes | OpenAI API key for LLM answer generation |
| `DARWIN_BASE_DIR` | no | no | Override the base data directory |

## Tools (11)

- `search_darwin`
- `search_lists`
- `get_search_results`
- `get_search_result_by_id`
- `get_schema`
- `run_pipeline`
- `purge_artifacts`
- `create_store`
- `list_files`
- `file_status`
- `get_logs`

## FAQ

### Does Darwin RAG work offline?

Yes, after an initial internet connection is used to download models during setup, all search and indexing operations run fully offline.

### Is my data private?

Yes, Darwin is designed to be local-first, meaning your documents stay on your own machine rather than being sent to a cloud index.

### Which AI clients can I use with Darwin?

It works with standard MCP clients such as Claude Desktop and Cursor.