rabbitmq-mcp

RabbitMQ MCP server for management and AMQP operations, with OpenAPI-driven call_id execution and semantic search.

Features

• MCP tools: search_ids / get_id / call_id
• OpenAPI-driven full HTTP management API calls
• Semantic search (optional embeddings, keyword fallback when dependencies are missing)
• Request/response validation (Pydantic first, jsonschema fallback)
• Destructive operation safety guard (disabled by default)
• AMQP: publish/consume/ack/nack/reject

Requirements

• Python 3.12+
• RabbitMQ Management Plugin (Management API)

Install & Run

Using uvx

uvx rabbitmq-mcp \
  --host localhost \
  --port 15672 \
  --user admin \
  --password admin123 \
  --allow-destructive false

From source

uv run --project . rabbitmq-mcp \
  --host localhost \
  --port 15672 \
  --user admin \
  --password admin123 \
  --allow-destructive false

CLI Parameters

Required:

• --host
• --port
• --user
• --password
• --allow-destructive (true|false)

Optional (with defaults):

• --amqp-port (default 5672)
• --vhost (default /)
• --scheme (default http)
• --verify-ssl (default inferred from scheme)
• --timeout (default 10.0)
• --max-connections (default 20)
• --max-keepalive (default 10)
• --retries (default 1)
• --retry-backoff (default 0.2)
• --vhost-cache-ttl (default 60)

MCP Tools

• search_ids: Search operationId by keyword/semantics (pagination supported).
• get_id: Return input/output schema for the given operationId.
• call_id: Execute the operationId (HTTP/AMQP auto routing).

MCP Config Example

Example JSON config for MCP clients that accept command/args (e.g., Claude Desktop):

{
  "mcpServers": {
    "rabbitmq": {
      "command": "uvx",
      "args": [
        "rabbitmq-mcp",
        "--host",
        "localhost",
        "--port",
        "15672",
        "--user",
        "admin",
        "--password",
        "CHANGE_ME",
        "--allow-destructive",
        "false"
      ]
    }
  }
}

Generation & Validation Scripts

# Validate OpenAPI
uv run --project . validate-openapi

# Generate Pydantic schemas
uv run --project . generate-schemas

# Generate registry
uv run --project . python scripts/generate_registry.py

# Generate embeddings (requires vector extra)
uv run --project . --extra vector python scripts/generate_embeddings.py