Service Level Agreement (SLA)
Last updated: October 13, 2026
1. Scope
Applies to the FalaAI API services (/v1), under the subscribed plan.
2. Target availability (Uptime)
99.9% (three nines) monthly availability, measured at the edge of the service (FalaAI API) — equivalent to at most about 43 minutes and 49 seconds of downtime per month. This is the market standard for commercial B2B APIs and SaaS.
3. Measurement
Availability is measured per calendar month: (minutes in the month − unavailable minutes − scheduled maintenance minutes) ÷ (minutes in the month − scheduled maintenance minutes). A period is unavailable when the external monitor reports down (status 8/9). The error rate (HTTP 5xx) comes from the API's internal metrics (request logs, aggregated daily). Scheduled maintenance windows (item 6) and unavailability beyond the Processor's control (item 8) are not included.
4. Error rate
Less than 0.5% of valid requests may result in server errors (HTTP 5xx) per month. Client errors (4xx) are not counted.
5. Rate limiting
Each plan defines requests per minute (RPM) and concurrent requests. When the limit is exceeded, the API returns 429 — which does not constitute unavailability or an SLA breach.
6. Scheduled maintenance
Windows previously announced on the status page (preferably during low-usage hours). They do not penalize availability.
7. Service credits
Any service credit for a breach will be defined by legal on a case-by-case basis upon the event — no pre-set percentages in this document.
8. Exclusions
Force majeure, misuse of the service, Internet/third-party failures and unavailability beyond the Processor's control.
9. Status and contact
Real-time status on the status page. Questions: contato_falaai@action.tec.br.