Summary
Production SLA options are available for eligible managed deployments.
We do not publish a general availability percentage. Service levels are agreed individually in the order form for eligible deployments. Without an agreed SLA, no specific availability is guaranteed — see the Terms of Service.
An SLA agreed with us defines the following elements.
1. Availability definition
Which service is covered (for example a dedicated endpoint or a high-availability deployment), and when it counts as available: typically, when the endpoint accepts authenticated requests and returns responses that are not caused by a fault on our side. The definition also states how partial degradation is treated.
2. Measurement
How availability is measured — the monitoring method and location, the probe interval, the measurement period (usually a calendar month) and how the result is calculated and reported. Incidents and maintenance are published on our status page.
3. Maintenance windows
When planned maintenance may take place, how far in advance it is announced, and whether maintenance within the agreed window counts against availability. Emergency maintenance to address security issues is described separately.
4. Service credits
What happens if the commitment is missed: the credit applied to the affected service's monthly fee for each level of shortfall, how credits are claimed and within what period, the maximum credit per month, and how credits relate to other remedies.
5. Exclusions
Events that do not count against availability, typically including:
- announced maintenance within the agreed window;
- issues caused by the customer's application, network, configuration or misuse of API keys;
- exceeding agreed rate limits or capacity;
- suspension in accordance with the Terms of Service;
- failures of third-party networks outside our control;
- force majeure events;
- beta, preview or pilot features.
6. Support
Support channels depend on the plan. Response targets apply only where they are agreed in the order form.
| Support level | Included in | Channels | Response target |
|---|---|---|---|
| Standard | Developer, Growth, Dedicated Node, Developer Mac | Email, Documentation | Agreed per contract |
| Priority | Scale, Managed AI Node, Dedicated & CI/CD Mac | Email, Shared chat channel | Agreed per contract |
| Enterprise | High Availability and custom infrastructure | Email, Shared chat channel, Named engineer | Agreed per contract |
7. Eligibility
SLAs are available for eligible managed deployments, such as dedicated and high-availability configurations. A high-availability architecture can reduce the impact of a single node failure; its exact failover design is agreed per deployment. To discuss service levels for your deployment, contact [email protected].