Guardrails API Documentation
Introduction
Example section for showcasing Guardrails API endpoints
Welcome
This section provides an overview of the available Guardrails endpoints
Enkrypt AI Guardrails can be accessed via the following endpoints:
https://api.enkryptai.com/guardrails/
Guardrails endpoints
/detect
: Detects prompts injection, toxicity, keywords and NSFW content.
/pii
: Detects PII content and de-anonymizes it.
/hallucination
: Detects hallucination in model output, given context
/adherence
: Detects adherence to context
/relevancy
: Detects relevancy of the response to the question
/status
: Checks if the API is up and running.
/models
: Returns list of models loaded
/health
: Returns health status of the service
Authentication
All API endpoints are authenticated using API Keys and can be accessed from the User Dashboard.