Skip to main content

Generate your first stack

The fastest way to get started with better-openclaw is using the interactive CLI wizard:
npx create-better-openclaw@latest
1

Run the generator

Execute the command above. The CLI will guide you through service selection, configuration options, and deployment preferences.
The wizard includes automatic dependency resolution—if you select n8n, it will automatically add PostgreSQL for you.
2

Choose your services

Select from 94 available services across 21 categories:
  • AI platforms (Ollama, Open WebUI, Dify)
  • Databases (PostgreSQL, Redis, Qdrant)
  • Automation (n8n, Home Assistant)
  • Monitoring (Grafana, Prometheus)
  • And many more
Or pick a preset to get started instantly (see below).
3

Configure your stack

The wizard will ask about:
  • Reverse proxy (none, Caddy, or Traefik)
  • Domain name for auto-SSL
  • AI providers (OpenAI, Anthropic, Google, etc.)
  • GPU passthrough for AI services
  • Monitoring dashboards
Port conflicts are detected automatically and services are reassigned to available ports.
4

Launch your stack

Navigate to your generated directory and start all services:
cd my-openclaw-stack
docker compose up -d
Your entire OpenClaw infrastructure is now running!
Use docker compose ps to verify all services are healthy and running.

Use a preset for instant setup

Skip the wizard and generate a complete stack instantly with presets:
# Web research with vector memory and scraping
npx create-better-openclaw --preset researcher --yes

Available presets

PresetServicesMemoryUse case
minimalRedis~1 GBBasic caching and state management
creatorFFmpeg, Remotion, MinIO, Redis~2 GBVideo creation and processing
researcherQdrant, SearXNG, Browserless, Redis~2.5 GBWeb research with vector memory
devopsn8n, PostgreSQL, Redis, Uptime Kuma, Grafana, Prometheus~3 GBMonitoring and automation
ai-playgroundOllama, Open WebUI, Qdrant, LiteLLM, Redis~6 GBMulti-model AI experimentation
coding-teamClaude Code, Codex, Redis, PostgreSQL~3 GBAI coding agents with shared state
full-stackAll core services + all skill packs~8 GBComplete OpenClaw superstack
The full-stack preset requires at least 8GB RAM and significant disk space. Start with a smaller preset if you have limited resources.

Non-interactive mode for automation

Perfect for CI/CD pipelines and automation scripts:
npx create-better-openclaw \
  --services postgresql,redis,n8n,grafana \
  --proxy caddy \
  --domain example.com \
  --yes

Explore available services

List all 94 available services:
npx create-better-openclaw services list
Filter by category:
npx create-better-openclaw services list --category database
Services are grouped into 21 categories including AI platforms, databases, automation, monitoring, browser automation, and more.

View preset details

See exactly what’s included in a preset before generating:
npx create-better-openclaw presets info researcher
This shows:
  • All included services with descriptions
  • Skill packs included
  • Estimated memory usage
  • Service dependencies

Validate your stack

Check your configuration before deploying:
npx create-better-openclaw validate \
  --services postgresql,redis,n8n
The validator checks for:
  • Port conflicts
  • Missing dependencies
  • Resource limits
  • Configuration errors

Manage your stack

Once your stack is generated, you can manage it with these commands:
cd my-openclaw-stack
npx create-better-openclaw add grafana
Adding or removing services automatically regenerates your docker-compose.yml with proper dependency resolution.

Deploy to production

Deploy your stack directly to Dokploy or Coolify:
npx create-better-openclaw deploy
Follow the interactive prompts to select your provider, enter your instance URL, and provide an API key. For non-interactive deployment:
npx create-better-openclaw deploy \
  --provider dokploy \
  --url https://dokploy.example.com \
  --api-key your-api-key

Next steps

Installation guide

Learn about system requirements, dependencies, and advanced installation options

Service catalog

Explore all 94 available services and their configurations

Build docs developers (and LLMs) love