First login: the gateway token and approving your devices
Where the token lives, why a new device has to be approved, and how to rotate or revoke a token when a laptop goes missing.
Hosting / OpenClaw hosting
OpenClaw hosting without a server to maintain
OpenClaw is a self-hosted AI agent that talks to you through WhatsApp, Telegram or Slack and acts on your behalf — so it has to run 24/7. Here it is provisioned automatically in its own isolated container.
Provisioned on creation, nothing to installIsolated container per agentEU-Central, US-East, US-WestGateway token rotation built in
Three container sizes, one rule: RAM first
Renews at the same price — no surprise increases.
Agent 2 GB
$12.49/mo
$149.90 per year, or $14.99 month to month
- 1 OpenClaw instance
- 2 GB RAM
- 1 vCPU
- 20 GB NVMe
- Gateway token rotation
Most popular
Agent 4 GB
$24.99/mo
$299.90 per year, or $29.99 month to month
- 1 OpenClaw instance
- 4 GB RAM
- 2 vCPU
- 40 GB NVMe
- Gateway token rotation
Agent 8 GB
$49.99/mo
$599.90 per year, or $59.99 month to month
- 1 OpenClaw instance
- 8 GB RAM
- 4 vCPU
- 80 GB NVMe
- Gateway token rotation
Sizing follows OpenClaw's published guidance: 2 GB RAM is the floor for a text agent on one channel, 4 GB is the recommended setup, and browser automation wants 8 GB. LLM API usage (Anthropic, OpenAI or another provider) is billed by that provider and is not included.
Why a container, not a VPS
Nothing to install
The package provisions a clean OpenClaw instance the moment the container is created; the platform keeps it and the Node.js runtime updated.
Isolated by design
Each agent lives in its own container with its own CPU, RAM and I/O caps — the same isolation our hosting plans use.
Backed up and restorable
The workspace — memory, files, configuration — is in the platform backups.
Your data, your region
Conversations and memory stay in the container, in EU-Central, US-East or US-West.
The security checklist, done by the platform
An agent that can act for you can also act wrongly for you. The platform's OpenClaw toolkit and our hardening cover the list so you don't have to.
- –Gateway access is token-protected and the tokens rotate automatically.
- –Device authentication — approve each device by hand or let the platform do it, your choice.
- –Model provider keys live in the panel and can be swapped there, never pasted into chat.
- –Plugins only from the official registry, with a security audit run on a schedule.
- –Updates applied promptly — older versions of self-hosted agents stay vulnerable.
What every OpenClaw plan includes
Automatic provisioning, isolated container✓
Gateway token rotation, device auth✓
Updates, backups, one-click restore✓
Engineers on the ticket✓
Plus your LLM provider's API usage. Renews at the same price.
Set it up yourself — help is one ticket away
Open the knowledge baseStuck on a step? Open a ticket and an engineer will answer — not a bot.
Guide
Guide
Connect Telegram, WhatsApp or Slack
A bot token, an allowlist and one test message. The pairing defaults are strict on purpose — here is what each one does.
Guide
Add or rotate a model provider key
Where the key goes, how to switch the primary model, and why the spend limit belongs on the provider's side rather than yours.
Guide
Run the security audit and restore a backup
What `openclaw security audit` checks, which findings matter on a managed container, and how to roll the workspace back after an agent does the wrong thing.
OpenClaw hosting questions
Do I have to install anything?
No. The package provisions OpenClaw when your container is created. You log in, approve your device, add a model provider key and connect a channel — each step has a guide, and an engineer answers the ticket if you get stuck.
Do I get root or Docker?
No — this is a managed container, that's the point. If you need a whole server for your agents, the same OpenClaw toolkit runs as a dedicated package on a VPS or dedicated server of your own — ask for a quote.
Which AI model does it use?
Any provider’s API key — Anthropic, OpenAI and others. Local models need a GPU and far more RAM; these plans are sized for API-backed agents.
What does it cost beyond the plan?
Your LLM provider bills per token. A light personal agent typically costs less than the hosting; heavy browser-automation agents can cost more. Neither is included in our price.
Need a whole server for your agents?
The same OpenClaw toolkit runs as a dedicated package on a VPS or bare metal of your own — same panel, engineer-managed.