Use AI Guardrails by Zapier MCP to take action in your AI tools

Detect Personally Identifiable Information (ML-powered)

Scans text for personally identifiable information (PII) and returns pass/fail status with detected types. Select which PII types to check for, or leave all checked (default) to scan for all types.

Detect Prompt Attacks (LLM-powered)

Analyzes text for potential prompt attack attempts that could manipulate AI model behavior. Returns detection status and recommended action.

Sentiment Detection (ML-powered)

Analyzes text to determine its emotional tone (positive, negative, neutral, or mixed) with confidence scores for each sentiment category.

Detect Toxicity (LLM-powered)

Analyzes text for toxic or harmful content using AWS Bedrock. Returns whether toxicity was detected, the types found, and confidence scores.

Detect Toxicity (ML-powered)

Analyzes text for toxic or harmful content using AWS Comprehend. Returns toxicity score and detected label types.