Pick your plan. Your agent starts today.

Yes — and that's the point. Your LLM API key connects your OpenClaw agent directly to your chosen AI provider (Anthropic, OpenAI, DeepSeek, Gemini, or others). When your agent makes a call, it goes from your VM to your provider — not through Right Servers. We never see your prompts, your context, or your completions.

This means you control your AI spend at the source (no markup from us), you can switch models whenever you want without migrating anything, and your data flow is between you and the provider you've chosen and trust. Getting a key takes about 3 minutes on any major platform.

Any provider that OpenClaw supports natively. That includes Anthropic (Claude 3.5/3.7), OpenAI (GPT-4o and o-series), DeepSeek, Google Gemini, and others. As OpenClaw adds provider support, your RightClaw instance picks it up through managed updates. You can run different models for different agents or switch models without changing plans.

At the end of your first billing cycle, your plan renews at the standard rate — $9.99/month for Starter, $29.99/month for Managed. You'll receive a reminder before renewal. There's no price increase beyond that; the standard rate is the standard rate. If you want to cancel before the renewal, you can do so anytime from the portal with one click.

Upgrades happen immediately — your plan changes and you're prorated for the difference. Downgrades take effect at the next billing cycle. If you want to move from Managed to Enterprise, contact us and we'll handle the transition directly.

The infrastructure and security baseline is the same on both plans — private VMware VM, hardened config, managed OS patching, Canadian data residency. The difference is what we handle for you on top of that.

Managed adds: our curated and audited skill recipe library (an alternative to the ClawHub marketplace), ongoing threat monitoring, priority support, and a human who will actually help you configure your agent. Starter is for people who are comfortable with SSH and want the infrastructure managed without the hand-holding. Managed is for everyone else.

Your data is stored in Kitchener, Ontario, Canada. Right Servers operates within Canadian jurisdiction and complies with PIPEDA (Canada's federal private sector privacy law). Your data does not leave Canada, does not train any AI model, and is not shared with third parties. Your LLM API calls go from your Canadian VM directly to your chosen LLM provider — we don't intermediate those calls or log their content.

Yes. On all plans, you have the ability to install skills from ClawHub or from your own custom repositories. On Managed and Enterprise, we also give you access to our curated skill recipe library — skills we've audited for security and tested for functionality. If you're on Enterprise and want us to review a specific third-party skill before installation, we'll do that.

A note on ClawHub: the March 2026 security audit found credential-harvesting code in approximately 20% of ClawHub skills. We recommend using our curated library on Managed, and exercising caution with any unvetted ClawHub skill on any plan.

No. Starter and Managed are month-to-month. Cancel anytime from the portal. Enterprise may include a contract depending on the scope — we'll discuss that during your onboarding call. We don't believe in locking you in; we believe in earning your business every month.

Starter: Documentation, knowledge base, and email support. Response time within 1 business day.

Managed: Priority email support with faster response times, plus access to a team that's actually familiar with OpenClaw skill configuration. If your agent is misbehaving and you don't know why, we'll help you figure it out — not just point you at a doc.

Enterprise: Dedicated support contact, onboarding call, and an SLA. You'll have a person at Right Servers who knows your setup.

It means your VM ships with a specific set of configurations that close the most common attack vectors before you ever log in:

• UFW (Uncomplicated Firewall): Configured to allow only the ports your OpenClaw instance needs. Everything else is blocked by default.
• fail2ban: Automatically blocks IP addresses that make too many failed login attempts. Brute force becomes expensive for attackers.
• SSH key-only authentication: Password logins to your VM are disabled. If you don't have the key, you can't get in.
• Non-standard port assignment: We don't run services on their default ports — automated scanners looking for OpenClaw on default ports are looking in the wrong place.
• Minimal attack surface: No unnecessary services installed. The fewer open doors, the fewer doors that can be kicked in.

This is the baseline you'd get if you hired a competent DevOps engineer to harden your server before you started using it. We do it automatically, before you ever receive your credentials.