Something broke? Not on our watch.
Real engineers on call. Median first reply: 11 minutes.
How do I route traffic through Probe Guard?+
Change your SDK's base URL to your Guard endpoint (for example https://guard.lanbolab.pro/v1) and add the x-probe-policy header. Your existing provider key keeps working. Full setup takes about five minutes.
What latency should I expect?+
Median overhead is 18ms and p95 is under 30ms in managed regions. Self-hosted latency depends on your hardware; our Helm chart ships with sizing guidance.
Can I run Probe inside my VPC or air-gapped?+
Yes, on Enterprise. We provide a Helm chart for Kubernetes and an offline bundle with license file for air-gapped environments.
A legitimate request was blocked. What now?+
Open the request in Probe Trace, click "Mark as false positive", and Guard will learn from it within your tenant. You can also lower the policy threshold or add an allowlist rule in your policy file.
Does Probe store my prompts?+
Not unless you enable Trace. With Trace on, you control retention period, region, and whether we store full text or hashes only.
How do I rotate my Probe API key?+
Go to Settings → API keys → Rotate. Old keys stay valid for 24 hours so you can roll out without downtime.
How do refunds and plan changes work?+
Upgrades are prorated immediately. Downgrades take effect at the next billing cycle. Annual plans can be refunded pro-rata within the first 30 days.
All systems operational.
Tell us what happened.
The more detail, the faster we fix it. Request IDs from Trace are gold.
8 business hours
Production down? Choose Urgent and our on-call engineer gets paged immediately.
We're on it.
A copy of your ticket has been sent to . Reply to that email to add details.
- ✓Ticket created and encrypted
- ✓Routed to the on-call engineer for your product
- 3Engineer reviews and replies by email