# mozg

mozg gives agents MCP access to scored knowledge brains. It retrieves relevant notes, exposes coverage gaps, and carries working state into another session.

- MCP server · `type=mcp` · Claude Code, Codex, Cursor, GitHub Copilot, Windsurf (Devin Desktop), Cline, Zed, Gemini CLI · ★ 8
- Author: egorfedorov
- License: AGPL-3.0 · policy=link-only · description only; the file stays with the author
- Source: https://github.com/egorfedorov/mozg
- Updated: 2026-10-07
- HTML: https://affpapa.org/en/nevibecode/mcp/mozg

## What it does

mozg connects an agent to knowledge brains over MCP. A documentation source is split into atomic notes with embeddings, then checked against an exam of roughly 30 questions derived from the brain’s goal. Failed questions guide another focused read, while searches with no result can become new exam questions.

Retrieval runs on the server with hybrid search and a reranker, so the agent receives the passages it needs. It can list available brains, inspect their summaries, and use `brain_handoff` to carry working state into a later session. Reader corrections enter as attributed proposals and stay out of search until the owner reviews them. Notes are scanned for leaked credentials, personal data, and prompt-injection language.

The hosted MCP endpoint can be added to Claude Code with one command. A self-hosted setup needs `ANTHROPIC_API_KEY` and `BETTER_AUTH_SECRET`; its first boot downloads about 2.2 GB of embedding weights. That download is the main setup constraint on a local deployment.

## Requirements
- Remote server: `https://mozg.sh/mcp`

## Install

### Claude Code (native platform)

**Platform CLI: remote server**
```bash
claude mcp add --transport http mozg https://mozg.sh/mcp
```

### Codex

**Platform CLI: remote server**
```bash
codex mcp add mozg --url https://mozg.sh/mcp
```

### Cursor

**Config file**
File `.cursor/mcp.json`:
```json
{
    "mcpServers": {
        "mozg": {
            "url": "https://mozg.sh/mcp"
        }
    }
}
```

### GitHub Copilot

**Config file**
File `.vscode/mcp.json`:
```json
{
    "servers": {
        "mozg": {
            "type": "http",
            "url": "https://mozg.sh/mcp"
        }
    }
}
```

### Gemini CLI

**Platform CLI: remote server**
```bash
gemini mcp add -t http mozg https://mozg.sh/mcp
```

## When to use it
- search a connected knowledge brain
- measure documentation coverage with an exam
- surface gaps after zero-result searches
- carry agent state into a later session
- submit reviewed knowledge corrections as proposals

## Example prompts
- Search the connected mozg brain for the Expo setup and include the source.
- List the available brains and give me a short brief for each.
- Hand the current working context to the next session with `brain_handoff`.

## Similar
- [REA](https://affpapa.org/en/nevibecode/mcp/rea.md) — MCP server · ★ 392262
- [GitHub](https://affpapa.org/en/nevibecode/mcp/github.md) — MCP server · ★ 33989
- [World Monitor](https://affpapa.org/en/nevibecode/mcp/world-monitor.md) — MCP server · ★ 89609
- [Netdata](https://affpapa.org/en/nevibecode/mcp/netdata.md) — MCP server · ★ 81163
- [Front-End Checklist](https://affpapa.org/en/nevibecode/mcp/front-end-checklist.md) — MCP server · ★ 74569
- [Scrapling MCP Server](https://affpapa.org/en/nevibecode/mcp/scrapling-mcp-server.md) — MCP server · ★ 89899

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md