Service Level Agreement

This Service Level Agreement ("SLA") is entered into between Infinized B.V., a company registered in the Netherlands (KVK: 88526461, VAT: NL864668788B01), hereinafter referred to as "Infinized," and the customer ("Customer") who has subscribed to the Infinized AI Control Plane ("Platform").

This SLA applies to the core Infinized Platform services, including the Agent Builder, Model Gateway, Governance Engine, Runtime Manager, and associated APIs. It forms part of the Infinized Terms of Service and governs the service availability commitments, remedies, and support obligations provided by Infinized.

This SLA applies to customers on paid plans (Growth, Scale, and Enterprise). Customers on the Starter plan receive community-level support and are not covered by the uptime commitments or service credit provisions described herein.

Infinized commits to the following monthly uptime targets for the Platform, measured as a percentage of total minutes in a calendar month:

Monthly Uptime Percentage is calculated using the following formula:

Where:

• Total Minutes means the total number of minutes in the applicable calendar month.
• Downtime Minutes means the total number of minutes during which the Platform's core services are materially unavailable, excluding Excluded Downtime (as defined in Section 7).

The following periods are excluded from Downtime calculations:

• Scheduled maintenance windows (see Section 4).
• Force majeure events, including natural disasters, wars, government actions, and pandemic-related disruptions.
• Downtime caused by the Customer's actions, including misconfiguration, excessive API usage beyond rate limits, or misuse of the Platform.
• Outages of third-party AI model providers (see Section 10).
• DNS or network issues outside of Infinized's infrastructure.

Infinized performs regular maintenance to ensure Platform reliability, security, and performance. Scheduled maintenance is excluded from uptime calculations.

4.1 Maintenance Windows

Standard maintenance windows are:

• Primary window: Tuesdays and Thursdays, 02:00 - 06:00 CET.
• Extended window: First Sunday of each month, 00:00 - 08:00 CET (for major upgrades).

4.2 Advance Notice

• Standard maintenance: At least 72 hours advance notice via email and the Infinized status page.
• Emergency maintenance: As much notice as reasonably practicable, with real-time updates on the status page. Emergency maintenance is limited to critical security patches and urgent stability fixes.

4.3 Zero-Downtime Deployments

Infinized employs rolling deployments and blue-green deployment strategies to minimize service interruption. Most maintenance operations are performed without any user-facing downtime.

If Infinized fails to meet the applicable Monthly Uptime Target, the Customer is eligible for service credits as described below. Service credits are calculated as a percentage of the Customer's monthly subscription fee for the affected month.

To receive a service credit, the Customer must submit a request within thirty (30) calendar days of the end of the month in which the downtime occurred. Requests must include:

• The Customer's organization name and account identifier.
• The dates and times of the claimed downtime.
• A description of the impact on the Customer's operations.
• Any relevant logs, error messages, or screenshots.

Credit requests should be submitted to contact@infinized.com with the subject line "SLA Credit Request."

Infinized will review the request and respond within ten (10) business days. Infinized's monitoring data shall be the authoritative source for determining uptime and downtime.

This SLA does not apply to, and no service credits will be issued for, downtime or performance degradation resulting from:

• Customer-caused issues: Misconfiguration of agents, policies, or API integrations; exceeding documented rate limits; or any misuse of the Platform in violation of the Terms of Service.
• Third-party AI provider outages: Unavailability or degraded performance of upstream AI model providers, including but not limited to OpenAI, Anthropic, Mistral, and Google. See Section 10 for details.
• Force majeure: Events beyond Infinized's reasonable control, including natural disasters, wars, terrorism, government actions, pandemics, power outages, or widespread internet failures.
• Beta and preview features: Any feature, service, or API endpoint explicitly designated as "beta," "preview," "experimental," or "early access."
• Scheduled maintenance: Downtime during announced maintenance windows as described in Section 4.
• Customer's infrastructure: Issues with the Customer's own network, hardware, software, or third-party services not provided by Infinized.
• Abuse or security incidents: Downtime resulting from DDoS attacks, unauthorized access attempts, or other security incidents originating from or targeting the Customer's account.

Infinized provides tiered support based on the Customer's subscription plan. Response times refer to the initial acknowledgment of a support request, not resolution time.

Infinized is committed to transparent and timely communication during service incidents.

9.1 Status Page

Infinized maintains a public status page that provides real-time information about Platform availability, ongoing incidents, and scheduled maintenance. Customers are encouraged to subscribe to status page notifications.

9.2 Incident Updates

During an active incident, Infinized will provide updates at the following intervals:

• Critical incidents: Updates every 30 minutes until resolution.
• Major incidents: Updates every 60 minutes until resolution.
• Minor incidents: Updates at the time of detection and resolution.

9.3 Post-Incident Reports

For Critical and Major incidents, Infinized will publish a post-incident report ("PIR") within five (5) business days of resolution. The PIR will include a timeline of events, root cause analysis, impact assessment, and corrective actions taken to prevent recurrence.

The Infinized Platform routes requests to third-party AI model providers, including but not limited to OpenAI, Anthropic, Mistral, and Google. These providers operate independently of Infinized and maintain their own infrastructure and service level agreements.

Infinized takes the following measures to mitigate the impact of third-party provider outages:

• Automatic failover: The Model Gateway supports configurable fallback chains, allowing requests to be automatically routed to alternative providers when the primary provider is unavailable.
• Provider status monitoring: Infinized monitors the health of connected providers and reflects their status on the Infinized status page.
• Transparent reporting: Incidents caused by third-party providers are clearly distinguished from Infinized Platform incidents in all reporting.

11.1 Monitoring

Infinized employs comprehensive monitoring across all Platform components, including:

• Synthetic health checks running at 60-second intervals from multiple geographic locations.
• Real-time alerting for latency spikes, error rate increases, and resource utilization thresholds.
• End-to-end transaction monitoring for critical user flows (API calls, agent executions, model gateway routing).
• Infrastructure-level monitoring for compute, storage, and network resources.

11.2 Reporting

Infinized provides the following availability reporting:

• Public status page: Real-time availability information accessible to all customers.
• Monthly uptime reports: Available to Scale and Enterprise customers upon request, detailing uptime percentage, incident summaries, and maintenance windows.
• Quarterly business reviews: Available to Enterprise customers, including detailed performance metrics, trend analysis, and capacity planning.