Guardrails & Content Policies

Define custom rules that control what your AI can and cannot do.

Features

Topic Restrictions

Block or flag requests about specific topics. Keep AI conversations on-brand and within scope.

Output Validation

Ensure model responses match expected formats (JSON schema, word limits, required fields).

Banned Phrases

Block specific words or phrases in both requests and responses. Useful for brand safety and compliance.

Custom Rules

Write custom guardrail rules using a simple configuration. Combine conditions with AND/OR logic.