RaiConfig interface

Package:

@azure/ai-projects

Configuration for Responsible AI (RAI) content filtering and safety features.

Properties

invocations_moderation	Author-declared configuration telling the platform where user/agent text lives in the agent-defined invocations request/response bodies, so content-safety guardrails can extract and moderate it. Optional; a rai_config without it leaves the invocations path without content-safety moderation.
rai_policy_name	The name of the RAI policy to apply.

Property Details

invocations_moderation

Author-declared configuration telling the platform where user/agent text lives in the agent-defined invocations request/response bodies, so content-safety guardrails can extract and moderate it. Optional; a rai_config without it leaves the invocations path without content-safety moderation.

invocations_moderation?: RaiInvocationModeration

Property Value

RaiInvocationModeration

rai_policy_name

The name of the RAI policy to apply.

rai_policy_name: string

Property Value

string