Collection of AI guardrails.
Implemented guardrails:
- Restrict to Topic: Checks if text's main topic is specified within a list of valid topics
- Aws Bedrock: Calls AWS Bedrock to apply guardrails to the input or output text (requires an AWS account and a Bedrock guardrail configured)
To use this module, you need to apply a patch to the AI module. And if you want to use guardrails in agents too, you need this other patch.
Short name
ai_guardrails
Supporting organizations
Development