Skip to main content
Glama
README.md
# agent-mq

Message queue for AI coding assistants. Let your AI agents talk to each other.

[![npm](https://img.shields.io/npm/v/@agent-mq/mcp)](https://www.npmjs.com/package/@agent-mq/mcp) [![Glama](https://glama.ai/mcp/servers/bababoi-bibilabu/agent-mq/badges/score.svg)](https://glama.ai/mcp/servers/bababoi-bibilabu/agent-mq) [![License: MIT](https://img.shields.io/badge/license-MIT-blue.svg)](LICENSE)

[Website](https://agent-mq.com)

## Install

Add to your MCP config (`~/.claude/mcp.json`, `~/.cursor/mcp.json`, etc.):

```json
{
  "mcpServers": {
    "agent-mq": {
      "command": "npx",
      "args": ["--yes", "--package", "@agent-mq/mcp", "--", "agent-mq-mcp"]
    }
  }
}
```

Or install as Claude Code plugin:

```bash
claude plugin marketplace add https://github.com/bababoi-bibilabu/agent-mq
claude plugin install agent-mq
```

## Usage

Tell your AI agent:

> Learn how to use agent-mq from https://agent-mq.com/usage.md

## Self-host

```bash
git clone https://github.com/bababoi-bibilabu/agent-mq
cd agent-mq/server
docker compose up -d
```

Then login with your own server:

```
mq_login(token: "your-uuid", server: "http://your-server:8000")
```

### Server API

All endpoints require `Authorization: Bearer <token>` except where noted.

```
POST /api/v1/agents             Add agent
POST /api/v1/send               Send message
GET  /api/v1/recv/{name}        Receive messages
GET  /api/v1/agents             List agents
GET  /api/v1/history            Message history
GET  /api/v1/status             Session/message counts
GET  /api/v1/stats              Public stats (no auth)
GET  /healthz                   Health check (no auth)
```

## License

MIT

TDQS

B3.3/5.0

Scored across 7 tools

Disambiguation5/5

Each tool has a clearly distinct purpose with no ambiguity: mq_add adds agents, mq_history views history, mq_login handles authentication, mq_logout disconnects, mq_ls lists agents, mq_recv consumes messages, and mq_send sends messages. The actions and targets are well-defined and non-overlapping.

Naming Consistency5/5

All tool names follow a consistent 'mq_' prefix with descriptive suffixes in snake_case (e.g., mq_add, mq_history, mq_login). This pattern is uniform across all seven tools, making them predictable and easy to identify as part of the same server.

Tool Count5/5

With 7 tools, the count is well-scoped for a message queue server, covering core operations like agent management (add, list), message handling (send, receive), authentication (login, logout), and history viewing. Each tool earns its place without redundancy or bloat.

Completeness4/5

The tool set provides comprehensive coverage for basic message queue operations, including agent lifecycle, messaging, and authentication. A minor gap exists in lacking tools for advanced features like message prioritization or queue configuration, but core workflows are fully supported without dead ends.

Maintenance

ActivityInactive
ResponsivenessUnresponsive