Skip to main content
Project-wide settings that apply across all channels and environments. Review safety filters before going live. Access these settings from Behavior in the sidebar.

​
Safety filter defaults

Default content safety filters. Apply when a channel does not have its own overrides enabled.
Safety filters are configured on a per-channel basis. The defaults set here only apply when a channel’s safety filters are not explicitly enabled. Each channel can override these settings independently.
CategoryDescription
ViolenceControls filtering of violent content (Lenient → Strict)
HateControls filtering of hateful or discriminatory content (Lenient → Strict)
Sexual contentControls filtering of sexually explicit content (Lenient → Strict)
Self-harmControls filtering of self-harm related content (Lenient → Strict)
For the full reference – categories, severity behavior, language support, monitoring, and how filters interact with Guardrails – see Safety filters. Override the defaults per channel in Voice configuration and Chat configuration.

Safety filters

Full reference for content filter categories, severity levels, and per-channel overrides.

Safety dashboard

Monitor flagged conversations and safety metrics.

Chat configuration

Channel-specific chat safety and behavior settings.

Voice configuration

Channel-specific voice and safety settings.
Last modified on June 19, 2026