Pricing
Choose your capacity
Every plan includes unlimited local gpt-oss-120b inference, a fully isolated OpenClaw instance, and sovereign Sydney hosting. Pricing scales on CPU, memory, and storage, not on prompt volume. Pay yearly and save two months.
Lite
For personal agents and lighter workloads.
- 2 CPU cores
- 2 GB RAM
- 20 GB disk
- Unlimited local gpt-oss-120b
- WebChat + all channels
- Daily encrypted backups
- KVM-level isolation
Pro
For heavier single-user workloads and larger workspaces.
- 4 CPU cores
- 4 GB RAM
- 40 GB disk
- Unlimited local gpt-oss-120b
- WebChat + all channels
- Daily encrypted backups
- KVM-level isolation
Max
For large workspaces and the heaviest workloads.
- 8 CPU cores
- 16 GB RAM
- 100 GB disk
- Unlimited local gpt-oss-120b
- WebChat + all channels
- Daily encrypted backups
- KVM-level isolation
All plans include
The same trust model on every tier
Unlimited local inference
The default model path runs on Clawd.au-managed infrastructure in Sydney.
KVM isolation
Dedicated microVM per tenant via Kata Containers and Cloud Hypervisor.
Secure gateway
Authenticated ingress path with trusted-proxy enforcement before traffic reaches your tenant.
Prompt privacy
We keep usage and operational metrics, not a centralised prompt log.
Daily encrypted backups
Tenant data is backed up daily and stored off site in encrypted form.
WebChat + channels
Use the built-in web UI and connect supported messaging channels from one platform.
ClawHub skills
Full marketplace access. Install community skills with one click.
One-click upgrades
New OpenClaw versions appear in your dashboard with changelogs.
Australian hosting
Application services, model runtime, and tenant workloads stay in Sydney.
FAQ
Common questions
- Are there any token limits?
- No. Every plan includes unlimited local gpt-oss-120b inference. We price the service around infrastructure capacity rather than per-token billing.
- How can you offer unlimited inference?
- We run the default model stack ourselves on GPU capacity rented in Sydney. Because we operate the inference path directly, we are not reselling someone else's token meter.
- Do prompts leave Australia?
- Not by default. The default Clawd.au inference path stays on Australian infrastructure. Prompts only leave that path if you explicitly configure your own third-party provider.
- Do you log prompts?
- No central prompt logging. We collect service metrics and usage data for billing, capacity planning, and reliability, but not a platform-wide archive of prompt text.
- Can I bring my own API key?
- Yes. If you want to use an external model provider, you can bring your own API key. That provider relationship is optional and remains under your control.
- Where is the platform hosted?
- Clawd.au is hosted in Sydney, with production infrastructure in Equinix Sydney facilities and Australian-resident storage and runtime services.
- Can I upgrade or downgrade my plan?
- Yes. Change plans anytime from your dashboard billing page. Upgrades take effect immediately, downgrades at the end of your current billing period.
- What is KVM-level isolation?
- Each tenant runs inside a dedicated microVM powered by Kata Containers and Cloud Hypervisor. Your instance has its own kernel and memory space — real VM isolation, not shared containers.
- How do backups work?
- We take daily backups and store them off site in encrypted form so recovery does not depend on a single machine or facility.
- How do yearly plans work?
- Pay for 10 months upfront, get 12 months of service — saving 17%. Cancel anytime; we'll refund the unused portion pro-rata.
- Can I cancel anytime?
- Yes. Cancel from the billing page in your dashboard. Your instance stays active until the end of your current period.
Deploy your AI agent today
From $19/month. Live in under two minutes. Australian sovereign hosting.
Get started