Better OpenClaw exists to make self-hosted AI infrastructure accessible to everyone. We believe that deploying production-ready AI agent stacks should be as simple as running a single command.
- Zero to production in minutes: Users should be able to go from nothing to a running OpenClaw stack with minimal friction
- Interactive and guided: The CLI should guide users through decisions with helpful context
- Visual when helpful: The web UI provides a visual stack builder for those who prefer it
- Composable by default: Services should work well together with minimal configuration
- Security by default: Auto-generated secrets, secure defaults, TLS-ready reverse proxy
- Battle-tested configurations: Service definitions based on real-world production deployments
- Scalability considered: Architecture that can grow from development to production
- Monitoring included: Observability stack (Grafana, Prometheus, Loki) built-in
- Open and transparent: AGPL-3.0 license, open governance
- AI-friendly contributions: We welcome AI-assisted PRs with proper attribution
- Comprehensive documentation: Every service documented, every decision explained
- Skill-first thinking: OpenClaw skills are first-class citizens
- Docker Compose: The primary deployment target for simplicity
- Native support: Bare-metal recipes for performance-critical services
- Cloud-agnostic: Works on any VPS, homelab, or cloud provider
- Future-ready: Architecture that can support Kubernetes, Nomad, etc.
- ✅ Core service definitions (70+ services)
- ✅ Interactive CLI wizard
- ✅ REST API for programmatic generation
- ✅ Visual stack builder (web UI)
- 🚧 Mission Control (monitoring & management)
- 🚧 MCP server for AI agent integration
- 📋 Skill pack ecosystem
- 📋 Template marketplace
- 📋 Native installation recipes for all major services
- 📋 Multi-environment support (dev/staging/prod)
- 📋 Stack migration tools
- 📋 Health check dashboard
- 📋 Automated updates & backups
- 📋 Cost estimation & optimization
- 📋 Kubernetes manifests generation
- 📋 Terraform/OpenTofu providers
- 📋 Multi-node orchestration
- 📋 Advanced networking (service mesh, etc.)
- 📋 Compliance & audit tooling
- 📋 Marketplace for commercial services
Better OpenClaw is designed as a collection of focused packages:
@better-openclaw/core: Pure TypeScript schemas and generation logiccreate-better-openclaw: Interactive CLI for end users@better-openclaw/api: REST API for tooling integration@better-openclaw/web: Visual stack builder@better-openclaw/mission-control: Management & monitoring UI
Each package has a single, clear purpose. This makes the codebase maintainable and allows users to pick what they need.
We validate:
- Service definitions at build time (TypeScript + Zod)
- User selections in the CLI (dependency resolution)
- Generated configurations before writing (schema validation)
- Docker Compose files after generation (YAML validity)
This prevents broken stacks from ever being generated.
- Simple by default: The CLI asks only essential questions
- Advanced when needed: Power users can access all options via flags or config files
- Discoverable: Features are easy to find when you need them
While we write good docs, we prefer:
- Snapshot tests that ensure stable output
- Integration tests that validate real-world scenarios
- Type safety that prevents entire classes of bugs
Code that doesn't need documentation is better than well-documented code.
To stay focused, we explicitly exclude:
- Building Docker images: We reference existing images from Docker Hub, GitHub Container Registry, etc.
- Service development: We don't maintain the services themselves (Redis, PostgreSQL, etc.)
- Hosting: We provide the tooling, users provide the infrastructure
- GUI for everything: Some things are better done with code/config files
- Enterprise features behind paywall: Everything stays open source
We welcome contributions from everyone, including AI-assisted development. Our standards:
- Quality over quantity
- Tests required for new features
- Security consciousness
- Clear communication
- Technical decisions: Maintainers have final say, but community input is valued
- Service additions: Community can propose via PRs with proper testing
- Architecture changes: Discussed in GitHub Discussions first
Better OpenClaw is:
- Funded by community donations and sponsorships
- Maintained by volunteers passionate about self-hosted AI
- Committed to staying open source (AGPL-3.0)
We measure success by:
- Adoption: Number of stacks generated and deployed
- Contribution: PR count, issue engagement, skill submissions
- Stability: Low bug report rate, high uptime for generated stacks
- Satisfaction: User feedback, GitHub stars, community growth
Quality always beats quantity. We'd rather have 1,000 happy users than 100,000 frustrated ones.
This is a living document. As the project evolves, so will this vision. Last updated: March 2026.