Skip to main contentSkip to main content
Legal

Service Level Agreement

Our contractual commitments to availability, performance, and incident response.

Last Updated: March 15, 2026Effective: April 1, 2026

This Service Level Agreement (“SLA”) is entered into between Seismic Swift AI, Inc.(“Seismic Swift AI”) and the customer entity identified in the applicable Order Form (“Customer”) and forms part of the Master Services Agreement or Terms of Service governing Customer's use of the Service. This SLA sets forth Seismic Swift AI's availability commitments, performance targets, incident response obligations, and service credit remedies. In the event of any conflict between this SLA and the Master Services Agreement or Terms of Service, the Master Services Agreement or Terms of Service shall govern.

1.Definitions

1.1Availability” means the percentage of time during a given calendar month, excluding Scheduled Maintenance and Excluded Downtime, during which the Service is accessible and functional for Authorized Users.

1.2Downtime” means any continuous period of thirty (30) minutes or more during which the Service's core API endpoints fail to respond to health-check probes with a success status (HTTP 2xx), as measured by Seismic Swift AI's synthetic monitoring infrastructure.

1.3Excluded Downtime” means any period of unavailability attributable to: (a) Scheduled Maintenance notified in accordance with Section 8; (b) Force Majeure Events as defined in the Master Services Agreement; (c) actions or omissions of Customer or third parties not under Seismic Swift AI's control, including Customer's network, equipment, or software; (d) features designated as “Beta” or “Preview”; or (e) Customer's failure to comply with Documentation.

1.4Monthly Uptime Percentage” means: ((Total Minutes in Month − Downtime Minutes) / Total Minutes in Month) × 100.

1.5Service Credit” means the monetary credit applied to Customer's next invoice, calculated as a percentage of the Monthly Subscription Fee for the affected Service, as set forth in Section 4.

1.6Monthly Subscription Fee” means the recurring monthly fee paid or payable by Customer for the Service in the calendar month during which the SLA breach occurred.

2.Service Level Targets

2.1The following Service Level Targets (SLTs) apply to each subscription tier. Availability SLTs are measured on a calendar-month basis. API latency SLTs are measured at the Azure Front Door edge using synthetic monitoring probes executed at one-minute intervals from a minimum of three geographic probe locations.

MetricStarterProfessionalEnterpriseNotes
API Availability99.5%99.9%Calendar-month basis, excluding scheduled maintenance
API Latency P50< 200ms< 100ms< 75msMeasured at Azure Front Door ingress
API Latency P95< 1s< 500ms< 300msMeasured at Azure Front Door ingress
API Latency P99< 2s< 1sEnterprise only; measured at Azure Front Door ingress
Pipeline Completion99.0%99.5%Jobs completed without error within 2× estimated duration
Data Durability99.999%99.9999999%99.99999999999%Azure Blob ZRS (11 nines for Enterprise)
Incident Response (P1)Best effort1 hour15 minutesTime to first human acknowledgment
Incident Response (P2)Best effort4 hours30 minutesTime to first human acknowledgment
RPO4 hours1 hourMaximum data loss window in a DR event
RTO8 hours4 hoursTime to restore service in a DR event

2.2Starter tier subscriptions are provided on a commercially reasonable basis without contractual availability commitments. Service Credits do not apply to Starter tier subscriptions.

2.3The availability SLT for Professional tier (99.5%) allows for a maximum of approximately 219 minutes of Downtime per calendar month. The availability SLT for Enterprise tier (99.9%) allows for a maximum of approximately 43.8 minutes of Downtime per calendar month.

3.Measurement Methodology

3.1Synthetic Monitoring. Seismic Swift AI uses a synthetic monitoring infrastructure consisting of active health-check probes that execute HTTP GET requests against the Service's designated health endpoints (e.g., /health/ready) at one-minute intervals from a minimum of three geographically distributed Azure regions. A probe failure is defined as any HTTP response with a 5xx status code or a response timeout exceeding thirty (30) seconds.

3.2Incident Confirmation. Downtime is recorded only when at least two of the three monitoring probe locations simultaneously report a failed health check for a continuous period of thirty (30) minutes or more. Single-location failures are treated as probe errors and do not contribute to Downtime calculations.

3.3Measurement Period. Availability is calculated on a calendar-month basis (00:00:00 UTC on the first day of the month to 23:59:59 UTC on the last day of the month).

3.4Status Page. Real-time availability and historical uptime data are published at status.seismicswiftai.com. Monthly uptime reports are available in the Customer portal within five (5) business days following the end of each calendar month.

3.5Latency Measurement. API latency is measured as the server-side response time from receipt of the request at the Azure Front Door ingress point to transmission of the first byte of the response, excluding client-side network transit time. Latency percentiles are computed from all successful (HTTP 2xx) API requests in each calendar month.

4.Service Credits

4.1If Seismic Swift AI fails to meet the applicable Availability SLT in any calendar month, Customer is eligible to receive a Service Credit applied to the next invoice. Service Credits are calculated as a percentage of the Monthly Subscription Fee as follows:

Monthly Uptime RangeEnterprise CreditProfessional CreditDescription
99.9% – 99.5%10%N/A (within target)Marginal degradation
99.5% – 99.0%25%10%Moderate service degradation
99.0% – 95.0%50%25%Significant outage
Below 95.0%100%50%Severe / extended outage

4.2Service Credits are calculated based on actual Downtime minutes recorded by Seismic Swift AI's monitoring infrastructure. Credits for API latency SLT breaches are calculated as 10% of the Monthly Subscription Fee per 0.1 percentage point by which the monthly P95 latency exceeds the applicable target, up to a maximum of 25% of the Monthly Subscription Fee.

4.3The total Service Credits issued to Customer in any calendar month shall not exceed one hundred percent (100%) of the Monthly Subscription Fee for that month. Service Credits are not redeemable for cash, are non-transferable, and must be used within twelve (12) months of issuance.

4.4Service Credits are Customer's sole and exclusive remedy for any failure by Seismic Swift AI to meet the Service Level Targets set forth in this SLA, and shall not constitute an admission or evidence of any fault or liability on the part of Seismic Swift AI.

5.Exclusions

5.1This SLA and the Service Credits available hereunder do not apply to, and no Downtime shall be recorded for, any period of unavailability attributable to any of the following:

  • Scheduled Maintenance notified in accordance with Section 8 of this SLA;
  • Emergency maintenance required to address an imminent security threat, data integrity risk, or regulatory obligation, provided Seismic Swift AI provides notice within twenty-four (24) hours;
  • Force Majeure Events as defined in the governing agreement (including acts of God, natural disasters, war, terrorism, epidemic, pandemic, governmental action, or internet backbone failures beyond Seismic Swift AI's reasonable control);
  • Actions or omissions of Customer or its Authorized Users, including: misuse of the Service; invalid API calls; Customer-side network or equipment failures; Customer-modified configurations or integrations not supported by Seismic Swift AI;
  • Third-party services not under Seismic Swift AI's control (e.g., Customer's identity provider, third-party DNS providers, or Customer-initiated third-party integrations);
  • Features or APIs labeled as “Beta,” “Preview,” “Experimental,” or “Alpha” in the Documentation;
  • Starter tier subscriptions;
  • Free trials or evaluation accounts.

5.2In any month in which the total Excluded Downtime exceeds fifty percent (50%) of total Downtime, Seismic Swift AI reserves the right to reasonably recalculate the Monthly Uptime Percentage to exclude such Excluded Downtime, and shall provide Customer with written documentation supporting any such recalculation.

6.Claim Process

6.1To receive a Service Credit, Customer must submit a written credit request to support@seismicswiftai.com with the subject line “SLA Credit Request — [Month Year]” within thirty (30) days of the end of the calendar month in which the claimed SLA breach occurred. Claims submitted outside this window will not be honored.

6.2The credit request must include: (a) the date(s) and duration(s) of the claimed Downtime; (b) a description of how the Downtime affected Customer's use of the Service; and (c) any supporting information (e.g., error logs, API response codes, screenshots).

6.3Seismic Swift AI will review the claim against its monitoring records and respond within fifteen (15) business days with either: (a) confirmation of the credit amount; or (b) a written explanation of why the claim does not qualify for a Service Credit.

6.4Disputed credit determinations may be escalated to Seismic Swift AI's VP of Engineering by written notice within ten (10) business days of the initial determination. Seismic Swift AI will provide a final written determination within fifteen (15) business days of escalation.

7.Disaster Recovery and Failover

7.1Seismic Swift AI maintains documented Disaster Recovery (DR) procedures designed to meet the RPO and RTO targets set forth in Section 2. DR procedures are tested at least semi-annually through planned DR drills, the results of which are available to Enterprise customers upon request under a non-disclosure agreement.

7.2Data Replication. All Customer Data is replicated across at least two Azure Availability Zones within the primary Azure region. For Enterprise customers who have enabled geo-redundancy, Customer Data is additionally replicated asynchronously to a secondary Azure region in the same geographic area (e.g., US East to US West, West Europe to North Europe).

7.3Automated Failover. In the event of a primary availability zone failure, the Service will automatically fail over to the secondary availability zone with a target recovery time of fifteen (15) minutes. In the event of a regional failure requiring geo-failover, the target recovery time is as specified in the RTO column of the SLT table in Section 2.

7.4Data Integrity During Failover. Seismic Swift AI employs write-ahead logging, transactional outbox patterns, and cryptographic audit chains to ensure data integrity during and after failover events. No Customer Data submitted before the RPO window shall be lost in a properly executed failover.

7.5Communication During DR Events. In the event of a DR activation, Seismic Swift AI will post status updates at status.seismicswiftai.com and send email notifications to Customer's designated contacts at minimum every sixty (60) minutes until service is fully restored.

8.Maintenance Windows

8.1Scheduled Maintenance. Seismic Swift AI performs routine maintenance during low-traffic windows. Standard Scheduled Maintenance windows are: Saturday 02:00 – 06:00 UTC and Sunday 02:00 – 06:00 UTC. Seismic Swift AI will provide at least seventy-two (72) hours' advance notice of any Scheduled Maintenance via: (a) email to Customer's designated contacts; and (b) posting on the status page.

8.2Enterprise Custom Windows. Enterprise customers may negotiate alternative Scheduled Maintenance windows with their account representative, subject to operational constraints. Custom window agreements must be confirmed in writing and reflected in the applicable Order Form.

8.3Emergency Maintenance. In exceptional circumstances where Seismic Swift AI must perform unscheduled emergency maintenance to protect the security or integrity of the Service or Customer Data, Seismic Swift AI will provide as much advance notice as practicable and will post real-time updates to the status page. Emergency maintenance does not qualify for Service Credits but shall be documented in a post-incident report provided to affected Enterprise customers within five (5) business days.

8.4Maintenance Exclusion. Time spent in Scheduled or Emergency Maintenance (announced with at least two (2) hours of notice) is excluded from Downtime calculations and does not qualify for Service Credits.

Questions about this SLA? Contact legal@seismicswiftai.com