diff --git a/en/docs/ai-gateway/ai-guardrails/overview.md b/en/docs/ai-gateway/ai-guardrails/overview.md index 14e8735ad3..a6c63e17bc 100644 --- a/en/docs/ai-gateway/ai-guardrails/overview.md +++ b/en/docs/ai-gateway/ai-guardrails/overview.md @@ -41,57 +41,57 @@ They act as intelligent filters that:
- +
Content Length Guardrail
Restricts the maximum length of LLM requests or responses.
 - +
Regex Guardrail
Validates LLM requests or responses using custom regular expressions.
 - +
JSON Schema Guardrail
Ensures LLM requests or responses match a specified JSON schema.
 - +
Sentence Count Guardrail
Limits the number of sentences in LLM requests or responses.
 - +
URL Guardrail
Validates URLs in LLM responses, useful to detect hallucinated URLs.
 - +
Word Count Guardrail
Restricts the number of words in LLM requests or responses.
 - +
PII Masking with Regex
Safeguard Personally Identifiable Information (PII) in LLM requests or responses by leveraging user-defined regular expressions.
 - +
Semantic Prompt Guardrail
Evaluate the semantic similarity of incoming prompts against a predefined list to allow or deny LLM requests.
 - +
Semantic Tool Filtering
Selects and ranks available API tools by semantic relevance to the user's query; includes evaluation metrics comparing embedding providers.
 - +
Azure Content Safety Content Moderation
Perform content moderation on LLM requests or responses using Azure Content Safety's Content Moderation API.
 - +
AWS Bedrock Guardrail
Enforce AI API Safeguards with AWS Bedrock Guardrails