Managed Services

Use managed local and integrated services in Msty Claw

Managed Services reduce local setup burden by handling install/start/health workflows for supported components. They are designed to shorten time-to-first-success on new machines.

What managed means in Claw

For supported services, Claw can guide:

  • Installation
  • Startup
  • Model pulls/downloads
  • Health checks and availability validation

The main value is operational consistency. Managed flows reduce setup drift between machines and make first-run troubleshooting faster.

Managed local AI flows

Current managed local stack includes:

  • Ollama
  • MLX (platform constraints apply)

Typical managed flow:

  1. Save provider configuration
  2. Use in-app setup/start actions
  3. Download/select supported models
  4. Verify health and model readiness

Run this flow fully before creating production agents. Most downstream execution issues are easier to avoid than to debug later.

1-click services

Claw also supports one-click service setup for workflow controls such as Data Protection, Web Search, and Airlock. Airlock is used for policy-based allowed/blocked website controls on agent network access and depends on container runtime support.

Use these services to reduce manual setup overhead while keeping configuration centralized and repeatable.

Container runtime support

Container-backed features use the available local runtime:

  • macOS: Docker, Podman, or OrbStack
  • Windows: Docker or Podman
  • Linux: Docker or Podman

On macOS, OrbStack support uses Docker-compatible commands. If OrbStack is running but Claw cannot detect it, verify Docker is using the OrbStack context before changing agent settings.

Platform notes

  • Ollama managed setup supports macOS, Windows, and Linux
  • MLX managed setup/model-management is limited to compatible Apple Silicon paths
  • Computer Use is available on macOS and Windows and uses a local helper that can be installed or updated from the Tools settings area

If your team is cross-platform, standardize a primary provider path first and treat alternatives as explicit exceptions.

Operational guidance

  • Use managed setup for first deployment on each machine
  • Validate model availability before agent rollout
  • Keep provider settings and health checks in regular maintenance routines

Re-check managed service health after updates or machine restarts to catch availability regressions early.

Use these settings with managed services:

  • Settings > Model Providers for model/runtime configuration
  • Settings > Tools for search/tool capabilities
  • Settings > Security for network policy and execution controls