Service availability
Our target for general availability is 99.9% monthly uptime for the gateway API on paid plans, with automatic failover between approved models.
Pre-launch. This page describes the controls the platform is being built with. Each is verified before general availability, and we update this page as they go live.
What availability means
The gateway is available when it accepts and routes valid requests. An outage at a model provider counts against us only when you have an approved fallback model configured and the gateway fails to use it.
Service credits for missed availability will be set out in the Terms of Service and order forms at launch.
Support response targets
| Severity | Example | First response target |
|---|---|---|
| Critical | Gateway down for all requests | 1 hour, 24/7 (Enterprise plans) |
| High | Major feature failing, no workaround | 4 business hours |
| Normal | Partial issue with a workaround | 1 business day |
| Low | Question or feature request | 2 business days |
Business hours are 09:00 to 18:00 Malaysia time (UTC+8), Monday to Friday, excluding Malaysian public holidays.
Incident communication
- A public status page will go live at launch.
- Confirmed incidents will be posted within 30 minutes.
- For major incidents, we will publish a summary of cause and fixes within 5 business days.
- Scheduled maintenance will be announced at least 48 hours ahead.