← Nexus Legal

SLA v1.0 · in effect since 2026-05-16

Service Level Agreement

Formal commitments from Nexus Legal by API key tier. It forms an integral part of the Terms of Service. Compensatory credits are automatically credited to the customer's balance at the close of each month in which any of the commitments is breached.

Standard

Uptime

99.5 %

Latency p50

≤ 3 s

Latency p95

≤ 12 s

Rate limit

10 rpm

Support

Email · response ≤ 3 business days

Credit-back

10 % of monthly consumption if uptime < 99.5 %

Premium

Uptime

99.9 %

Latency p50

≤ 2 s

Latency p95

≤ 8 s

Rate limit

200 rpm

Support

Email + shared Slack · response ≤ 4 business hours

Credit-back

25 % of monthly consumption if uptime < 99.9 %

Enterprise

Uptime

99.95 % (target) · 99.9 % guaranteed

Latency p50

≤ 1.5 s

Latency p95

≤ 6 s

Rate limit

1 000 rpm (configurable)

Support

Dedicated Slack + phone · response ≤ 1 business hour · named DRI

Credit-back

50 % of monthly consumption if uptime < 99.9 % · 100 % if < 99 %

Definitions

Monthly uptime: percentage of the calendar month during which the API responds with HTTP status codes < 500 to valid requests. Excluded from the calculation are: (a) maintenance windows announced with ≥ 48 h notice (maximum 4 h/month total), (b) incidents attributable to upstream sub-processors (OpenRouter, Anthropic, Supabase) when those sub-processors also breach their own SLA, (c) DDoS attacks or abusive traffic from the customer's own account.

Latency p50 / p95: response time of the `POST /api/v1/analyze` endpoint measured at the 50th / 95th percentile during the calendar month, for documents of up to 30 000 characters and `standard` mode. Analyses served by a failover route instead of the role's primary provider are excluded from the calculation, because their latency is set by the third party the request is redirected to. The latency commitments are not expressed in terms of any specific model or provider: the chain is provider-agnostic.

Rate limit: per-minute limit per API key. The customer may request a temporary increase via a support ticket; the effective limit configured in `api_keys.rate_limit_rpm` applies.

Credit-back: automatic compensation in the form of Nexus credits credited to the customer's balance at the close of the calendar month in which the commitment was breached. Calculated as a percentage of the billed consumption of the breached month.

Scheduled maintenance

  • Notification with ≥ 48 hours notice by email + banner in the Developer Portal.
  • Maximum 4 hours/month cumulative.
  • Preferred windows: Sundays 02:00-06:00 UTC.
  • Zero-downtime database migrations when technically feasible.

Incident reporting

The customer must report the incident within a maximum of 7 days from its occurrence to be entitled to automatic compensation. Report via:

Status page

Check real-time status at legal.nexusquantum.legal/status — historical monthly uptime, past incidents, and scheduled maintenance.

Limitations

  • The latency commitments apply only to well-formed requests within the tier's rate limit.
  • Uptime is measured at the global backend level; specific routes may have slightly different availability (e.g. `/api/v1/sandbox/analyze` is not included in the calculation because it does not consume credits).
  • Force majeure events (geopolitical causes that cut regional connectivity, blocks due to sanctions on sub-processors, etc.) temporarily suspend the commitments without generating credit-back.

Changes to the SLA

Any modification to the SLA is notified 60 days in advance to customers with a Premium/Enterprise tier. Changes beneficial to the customer (better uptime, lower latency, better support) may be applied immediately.

Does your organization need a custom SLA with additional clauses (specific data residency, dedicated DRI, integration with your internal status page, etc.)? Contact enterprise@nexusquantum.legal.