Service Level Agreement (SLA)
Last updated: 2026-06-07
1. Uptime commitment
We commit to 99.5% monthly uptime of the Service for Enterprise customers.
"Uptime" means the percentage of total minutes in a calendar month during which the Service is
available, measured as the success rate of synthetic checks against the POST /v1/detect endpoint executed every minute from at least three independent regions.
| Monthly uptime | Allowed downtime / month | Service credit |
|---|---|---|
| ≥ 99.5% | ≤ 3h 39m | None |
| ≥ 99.0% and < 99.5% | 3h 40m – 7h 18m | 10% of monthly fee |
| ≥ 95.0% and < 99.0% | 7h 19m – 36h 30m | 25% of monthly fee |
| < 95.0% | > 36h 30m | 50% of monthly fee |
Service credits are issued as a credit toward your next invoice. They are not paid in cash and are not transferable. They are the sole and exclusive remedy for failure to meet the uptime commitment.
2. Exclusions from downtime
The following time is excluded from the uptime calculation:
- Scheduled maintenance announced at least 48 hours in advance, capped at 4 hours per month.
- Emergency maintenance for security or stability, with as much notice as practicable.
- Force majeure events outside our reasonable control (eg, internet-backbone failures, sub-processor outages beyond our control, government action).
- Failures caused by Customer's own systems, integrations, throttling, misconfiguration, or violations of the AUP.
- Failures caused by Customer's exceeding documented rate limits or contracted volume.
- Beta features and Customer-specific preview environments.
- Issues on third-party networks between Customer's user and our edge that we cannot remediate.
3. Performance targets
- API latency (p50): ≤ 1.5 s for documents up to 5,000 words on the standard pipeline.
- API latency (p95): ≤ 4.0 s for documents up to 5,000 words on the standard pipeline.
- Dashboard time-to-interactive (p50): ≤ 2.0 s on a 4G connection.
- Long-document & deep-scan endpoints: targets defined per workload in the order form.
Performance targets are best-effort. Failure to meet a performance target is not a breach of the uptime commitment and does not by itself trigger service credits.
4. Support response times
| Severity | Definition | First response | Status updates |
|---|---|---|---|
| S1 — Critical | Production Service unavailable; no workaround. | 1 hour, 24×7 | Every 2 hours until mitigated |
| S2 — High | Major function impaired; workaround available. | 4 hours, business hours | Daily until resolved |
| S3 — Normal | Minor function impaired or question. | 1 business day | As needed |
| S4 — Low | Cosmetic / feature request. | 3 business days | As needed |
Business hours: 09:00–18:00 UTC, Monday–Friday excluding US federal holidays. S1 tickets are 24×7. Severity is initially declared by Customer and confirmed by us within first response.
5. Status page & communication
- Public status page at
status.deepaidetector.com(planned, in progress). - Incident updates posted on status page, RSS, and to Enterprise subscribers' designated email.
- Post-incident summary within 5 business days of S1 incidents.
- Maintenance announced at least 48 hours in advance for non-emergency work.
6. How to claim a service credit
- Email sla@deepaidetector.com within 30 days of the end of the affected month.
- Include: account, monthly fee, affected dates/times, supporting evidence (eg, your own monitoring) sufficient to identify the downtime.
- We respond within 10 business days, apply the credit to the next invoice if accepted, or explain the basis for declining.
7. Maximum service-credit cap
Total service credits in any calendar month will not exceed 50% of that month's fees. Service credits are the sole and exclusive remedy for failure to meet this SLA; nothing in this SLA expands liability beyond Terms §13.
8. Changes
This SLA may be updated. Material changes to active Enterprise commitments require written agreement on a renewal or order form.
9. Contact
For SLA matters: sla@deepaidetector.com. For sales / Enterprise order forms: enterprise@deepaidetector.com.