What is included with an instance?
A dedicated OpenClaw Gateway container, isolated network, persistent volume, plan-based CPU and RAM allocation, and a managed HTTPS hostname.
Questions, answered.
Practical details about how isolated OpenClaw hosting works.
Do you run multiple customers in one Gateway?
No. Each customer gets a separate OpenClaw Gateway container, volume, configuration, and credentials.
How does provisioning begin?
The control plane provisions only after Stripe sends a valid signed event confirming the payment. The API does not trust browser payment status.
Can I bring my own AI provider?
Yes. Configure OpenAI, Anthropic, Gemini, OpenRouter, an OpenAI-compatible endpoint, or Ollama from the instance settings.
Can I change plans later?
Plan resources are snapshotted when the subscription is created. Contact support to request a change; the existing resource allocation does not silently shift when a plan is edited.
Where does my instance run?
The scheduler selects an online Docker worker with enough recorded CPU, RAM, storage, and container capacity. Node regions are visible to administrators.
Can I cancel and keep my data?
The platform retains instance data according to its configured retention policy. Contact support before your retention window ends if you need an export.
Are model provider keys exposed to OpenClaw Cloud staff?
Keys are encrypted in the control-plane database and are never returned by API responses. The Docker worker must make the key available to the customer gateway at runtime.