Thanks — I’ll answer both questions clearly. • Under what circumstances do you actively monitor usage of the Services? ◦ LiveKit does not actively monitor customer usage by default. LiveKit may elect to monitor or investigate usage when there are suspected violations of the Acceptable Use Policy or Terms of Service (for example, spam, fraud, unauthorized access, or other prohibited activity) or when required to comply with laws or regulatory obligations. Monitoring can also occur to remove or mitigate content that violates policies or to respond to lawful requests from authorities. • Does monitoring include access to call audio content, or is it limited to metadata and abuse signals? ◦ Monitoring is generally limited to telemetry, metadata, and abuse signals (session-level logs, metrics, call detail records, prompts/responses used for service operation, and other observability data). Audio processing required for service features (for example VAD or transcription used to deliver the service) is performed as needed, but LiveKit does not retain inference inputs/outputs or use customer content for model training. The sources do not state that routine human review of raw call audio occurs; audio is processed for operational/compliance purposes and monitoring focuses on metadata/abuse signals. If you need, I can: • Point to the specific policy passages for each point. • Clarify how audio is handled in a particular feature (transcription, agent-hosting, or recording). Which one do you want more detail on? Sources: Acceptable Use Policy | LiveKit | Terms of Service | LiveKit | Capturing the First Few Seconds of a Call for Processing | Privacy Policy | LiveKit, with interactive elements