99.9% monthly uptime for the Standard Compute API, with service credits if we miss it. Included in Enterprise and in the Uptime SLA add-on.
This Service Level Agreement ("SLA") forms part of the agreement between Standard Compute AS, org. no. 938 498 806, Oslo, Norway ("Standard Compute") and customers on the Enterprise plan or with the Uptime SLA add-on ("Customer"). See pricing for plans and the status page for live status.
1.1Standard Compute will make the API available with a Monthly Uptime Percentage of at least 99.9% in each calendar month.
1.2The API means the Standard Compute endpoints at https://api.stdcmpt.com/v1, including chat completions, completions, responses, messages and models. The website and dashboard are not covered.
2.1Monitoring. External checks call https://api.stdcmpt.com/v1/models every 5 minutes from Europe, Iowa and Virginia. A 5-minute interval counts as down when the API fails in at least two of the three regions.
2.2Downtime interval. A 5-minute interval in which the check does not receive a successful response within 10 seconds in at least two of the three regions.
2.3Monthly Uptime Percentage. The number of measured 5-minute intervals in the calendar month (UTC) that are not Downtime intervals, divided by all measured intervals, as a percentage. Excluded periods under section 3 are left out of both numbers.
2.4Your evidence. If you believe the API was unavailable when our checks passed, send request IDs and times with your claim. We review them in good faith and count confirmed unavailability as Downtime.
3.1Scheduled maintenance announced on the status page at least 48 hours in advance, up to 4 hours per calendar month.
3.2Errors caused by the Customer or its users, including invalid requests, revoked or missing API keys, reached budgets or spend limits, rate limits, and region or provider rules that leave no eligible provider for a request.
3.3Outages of an upstream model provider when no eligible alternative provider exists for the requested model. The API itself stays available and returns an error in that case.
3.4Accounts suspended for non-payment or under the acceptable use policy, and features labelled beta or preview.
3.5Events beyond Standard Compute's reasonable control, such as internet failures outside our network, acts of government or natural disasters.
4.1If the Monthly Uptime Percentage falls below the commitment, the Customer receives a service credit on that month's Enterprise fee or Uptime SLA add-on fee, as shown in the table above. The lowest band reached applies; credits do not add up.
4.2Credits are applied to the next invoice. They are not paid out in cash, and the credit for one month never exceeds 50% of that month's fee.
4.3Service credits are the Customer's sole remedy for unavailability covered by this SLA.
5.1Email contact@standardcompute.com within 30 days after the end of the affected month, with the company name, the month and, if you have them, request IDs and times.
5.2We confirm the claim and the credit within 10 business days, using the measurements in section 2.
6.1Customers with the SLA see each month's uptime under Billing in the dashboard. Live status is on the status page.
6.2Standard Compute may update this SLA with 30 days' notice. Changes do not reduce the commitment during a paid period that has already started.