Service Level Agreement (planned)

Planned SLA terms.

2026-05 statusNo SLA framework is wired today — no service-credit issuance, no breach reporting, no status-page SLO tracking. The terms below describe the contract we plan to offer once the Team and Enterprise tiers ship with billing. The factory itself runs on a best-effort basis. See /factory/proof for real run history.

99.9% target on Team. 99.95% target on Enterprise. Best-effort on Hobby and Pro. Team and Enterprise are planned tiers (no billing wired). Live component status at /factory/status; automated 90-day uptime tracking is on the roadmap, so we do not yet publish a measured-uptime number.

This page is the published SLA. The contract version (Enterprise MSA) is the controlling document.

Per-tier SLA

TierUptimeSupport responseService credits24/7 pager
Hobby (free) — liveBest-effort (no formal SLA)Best-effortNoneNo
Pro — plannedBest-effort (no formal SLA)Email within 4 business hours (target)NoneNo
Team — planned99.9% monthly target (Service Credits below)Email within 1 business hour (target)10% / 25% / 50% (see below)No
Enterprise — planned99.95% monthly target (custom contract)Email within 30 min, 24/7 (target)Custom (negotiated in MSA)Yes — direct pager rotation

Service credit schedule (Team)

If we breach the Team-tier SLA in a given calendar month, you receive automatic credits applied to your next invoice. Credits cap at 50% of the monthly fee. Enterprise customers can negotiate higher caps in their MSA.

Monthly uptimeService credit
99.0% – 99.9%10% of monthly fee
95.0% – 99.0%25% of monthly fee
below 95.0%50% of monthly fee

We compute uptime monthly from our internal observability + our public /factory/status page. Credits are auto-applied; you do not need to file a claim. Service credits are the sole remedy for an SLA breach.

What’s covered

Operator at /admin/factory
Page must render within 3 seconds at p95.
Public API at /api/factory/*
All endpoints listed in /api/factory/openapi. p95 response < 3s for sync endpoints; SSE streams measured separately.
Linux bridge (Runner-1)
Build execution. Excludes time spent in upstream AI provider latency.
Windows bridge
Build execution for desktop surface. In-progress as of 2026-05; SLA activates when generally available.
AI provider latency / errors
Failures attributable to Anthropic, OpenAI, Google, or Mistral are NOT counted against our SLA. We fall back to alternate models automatically when available.
Scheduled maintenance
Announced ≥7 days in advance, never more than 30 min/month, never during US business hours. Not counted against SLA.
Force majeure
Cloud provider region failure, DDoS attacks, government action.

Need a custom SLA?

Enterprise contracts can carry custom uptime targets, credit schedules, RPO/RTO commitments, and named-account pager rotation. Email [email protected].

© 2026 OOretz Factory · SLA published; MSA controls