Skip to content

Install and manage OpenClaw

Self-hosted AI assistant that requires separate model and channel configuration.

Protect your data firstKeep a current off-server backup before reinstalling, updating, changing firewall rules, or running an application installer.

Before installation

Use a supported Linux image and a plan that meets the application requirements. Back up any existing workload. Confirm DNS, free disk space, and required ports before approving the installer.

Install from the client panel

  1. Open the purchased VPS and choose Install application.
  2. Select OpenClaw and review compatibility, estimated time, and the exact plan.
  3. Confirm the server identity and backup warning.
  4. Follow installation progress from the task page. Do not run the installer twice.

The VPS can show Active while application setup is still finishing. Wait for Ready or review the reported result before using it.

First setup

Installation prepares the software; it does not provide a model account or messaging-channel credentials. Complete onboarding with your own approved model provider, restrict public access, and grant only the tools the agent genuinely needs.

Troubleshooting

Check the installation result, VPS power state, DNS, required ports, free storage, memory, service or container status, and the first relevant log error. Preserve logs before retrying. If the task needs attention, open a support ticket instead of rebuilding the VPS.

Backup and recovery

Back up OpenClaw configuration, memory, approved skills, channel settings, and recovery information without exposing API keys. Autonomous tools can change files or external systems, so begin with minimum permissions.

Official documentation

Review the current OpenClaw documentation before production changes.

Was this not enough?Include the service number, exact error, and time of the problem. Never send a password or private key.
Contact support