How to Start the Code-Graph-RAG Daemon: A Complete CLI Guide

Use the cgr daemon up command to start the Code-Graph-RAG daemon, which launches a Docker-Compose stack containing Memgraph and Qdrant via the StackManager class.

The Code-Graph-RAG daemon provides the backend infrastructure for graph-based retrieval-augmented generation operations. According to the vitali87/code-graph-rag source code, the daemon is orchestrated entirely through a CLI tool that wraps Docker-Compose operations, eliminating the need for manual container management.

Prerequisites for Starting the Code-Graph-RAG Daemon

Before launching the daemon, ensure your environment meets two requirements.

• Docker must be running on your system. The daemon relies on containerized services defined in the repository's docker-compose.yml.
• The package must be installed. Install via pip:

pip install code-graph-rag

Starting the Daemon with cgr daemon up

The primary command to start the Code-Graph-RAG daemon is cgr daemon up. This single command handles container creation, health checks, and service readiness.

cgr daemon up

Under the hood, this command performs several operations defined in the source code:

1. The CLI entry point in codebase_rag/cli.py registers the daemon command group and forwards calls to the stack CLI (lines 45-55).
2. The up sub-command in codebase_rag/stack/cli.py instantiates StackManager and invokes ensure_running() (lines 34-41).
3. The StackManager class in codebase_rag/stack/manager.py builds the Docker-Compose project, starts the Memgraph and Qdrant containers, and polls health endpoints until services are ready.

Verifying Daemon Status

Confirm the daemon is running with the status command:

cgr daemon status

This outputs the current state, service endpoints for Memgraph and Qdrant, and the path to the Compose file. The implementation calls StackManager.status() and warn_if_ports_are_public() in codebase_rag/stack/cli.py (lines 58-65).

Monitoring Daemon Logs

For troubleshooting or observability, view container logs through the CLI:

cgr daemon logs              # Show all services

cgr daemon logs -f           # Follow logs in real-time

cgr daemon logs -s memgraph  # Show only Memgraph logs

The logs command forwards to StackManager.logs() as implemented in codebase_rag/stack/cli.py (lines 80-92).

Stopping the Daemon

When you no longer need the services, shut down the stack cleanly:

cgr daemon down

This command triggers the Docker-Compose shutdown sequence through StackManager (lines 46-53 in codebase_rag/stack/cli.py).

Complete Command Reference

Command	Purpose
cgr daemon up	Starts the Docker-Compose stack and waits for healthy state
cgr daemon down	Stops and removes all containers
cgr daemon status	Displays running state and endpoints
cgr daemon restart	Restarts the entire stack
cgr daemon logs	Shows or follows container logs

Key Implementation Files

Understanding the source structure helps with debugging or extending the daemon:

• codebase_rag/cli.py — Top-level CLI registration, delegates to stack commands
• codebase_rag/stack/cli.py — Implements up, down, status, restart, and logs sub-commands
• codebase_rag/stack/manager.py — Core StackManager class handling Docker-Compose orchestration and health checks

Summary

• Install with pip install code-graph-rag and ensure Docker is running
• Start the Code-Graph-RAG daemon using cgr daemon up, which invokes StackManager.ensure_running()
• Monitor with cgr daemon status and cgr daemon logs commands
• Stop cleanly with cgr daemon down when finished
• The daemon stack comprises Memgraph (graph database) and Qdrant (vector database) containers orchestrated via Docker-Compose

Frequently Asked Questions

What services run inside the Code-Graph-RAG daemon?

The daemon manages two containerized services: Memgraph for graph storage and Qdrant for vector embeddings. Both are defined in the repository's docker-compose.yml and orchestrated by the StackManager class.

Why does cgr daemon up hang or timeout?

The command waits for health checks to pass. If Docker is not running, ports are already in use, or containers fail to start, ensure_running() will retry until timeout. Run cgr daemon logs to diagnose startup failures.

Can I run the daemon without the CLI?

While possible by invoking docker-compose directly on the repository's compose file, the CLI provides essential functionality including automated health checking, status reporting, and log aggregation through StackManager.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}