本服务面向非中国大陆地区 · ZetaAPI.AI 的网站与控制台主要面向非中国大陆市场,注册与开通功能在你当前地区不可用。

Service compensation claims (SLA)

Which incidents qualify for compensation, how the tiers are computed, how to file a claim and how long it takes. Currently claim-based, settled by manual review against the published algorithm.

Updated 2026-07-10

When a platform-side incident causes your requests to fail, you can claim compensation following this page. Compensation is currently claim-based: after you file, we review and settle by the published algorithm, crediting your account balance.

Scope (what counts toward SLA)

These three platform-side failure types count:

These do not count:

Even during large-scale upstream outages (force-majeure scenarios), in-scope 5xx / timeouts are still settled by the tiers above — we simply do not add extra compensation beyond the published rules (consistent with the Terms).

Compensation tiers (calendar month × SKU group)

Availability is computed per calendar month, per SKU group (e.g. Claude chat, GPT chat, Gemini chat):

Monthly availabilityCompensation (of that SKU’s monthly spend)
≥ 99%Not triggered
98% – 99%2%
95% – 98%5%
< 95%10%

Compensation for a single SKU is capped at 10% of that SKU’s monthly spend. Design background: Refund rules as product.

How to file

Email support@zetaapi.ai (or use the in-console chat) with:

  1. Your account email
  2. The affected time window (with timezone)
  3. The models or SKUs involved
  4. If available: error samples, x-request-id, failure-rate screenshots

Turnaround: we normally reply with the settlement result within 5 business days. Approved compensation is credited straight to your balance (credits never expire).

Appeal

If you disagree with the result, reply within 7 days and we will re-check the raw logs and give a final conclusion.