better-openclaw Skill

Use this skill to generate Docker Compose stacks via the better-openclaw API. The API returns complete docker-compose.yml, .env, reverse proxy configs, and OpenClaw skill files ready to deploy.

API Base URL

• Web API: https://better-openclaw.dev/api
• Standalone API: https://api.better-openclaw.dev/api/v1 (or configured API_URL)

Generate a Stack

POST to /generate (web) or /api/v1/generate (standalone):

curl -X POST "https://better-openclaw.dev/api/generate" \
  -H "Content-Type: application/json" \
  -d '{
    "projectName": "my-stack",
    "services": ["ollama", "qdrant", "n8n"],
    "skillPacks": ["ollama-local-llm", "qdrant-memory", "n8n-trigger"],
    "proxy": "caddy",
    "domain": "example.local",
    "monitoring": true
  }'

List Services

GET available services (standalone API only):

curl "https://api.better-openclaw.dev/api/v1/services"
curl "https://api.better-openclaw.dev/api/v1/services?category=ai"

List Skill Packs

curl "https://api.better-openclaw.dev/api/v1/skills"
curl "https://api.better-openclaw.dev/api/v1/skills?services=ollama,qdrant"

Validate Before Generation

Validate a stack configuration without generating files:

curl -X POST "https://api.better-openclaw.dev/api/v1/validate" \
  -H "Content-Type: application/json" \
  -d '{
    "services": ["ollama", "qdrant"],
    "skillPacks": ["ollama-local-llm"],
    "gpu": false
  }'

Generate Request Schema

Field	Type	Description
projectName	string	Output directory / project name
services	string[]	Service IDs (e.g. ollama, qdrant, n8n, minio)
skillPacks	string[]	Skill pack IDs (optional)
proxy	string	none, caddy, traefik
domain	string	Domain for auto-SSL (optional)
gpu	boolean	Enable GPU for AI services
monitoring	boolean	Add Grafana + Prometheus
platform	string	linux/amd64 or linux/arm64
deployment	string	local, vps, homelab

Common Service IDs

• AI: ollama, whisper
• Vector DB: qdrant, chromadb, weaviate
• Automation: n8n
• Storage: minio, redis, postgresql
• Proxy: caddy, traefik
• Media: ffmpeg, remotion

Tips for AI Agents

• Always include required dependencies: e.g. n8n adds postgresql automatically.
• Use validate before generate to catch port conflicts and config errors.
• The response contains a files object: keys are paths, values are file contents.
• Skill packs require their backing services (e.g. ollama-local-llm requires ollama).
• For homelab/VPS deployments, set proxy: "caddy" or traefik and provide domain.