Amazon Bedrock Guardrails

service

proprietary managed AWS service(no self-host)

features

denied topics, content filters, PII redaction, contextual grounding

source

closed

license

proprietary

Bedrock Guardrails is a closed AWS service used under AWS's service terms, with no published source or downloadable build. The ApplyGuardrail API can screen self-hosted models, but each check is still a call to the guardrail running at AWS.

• https://aws.amazon.com/bedrock/guardrails/ recorded 2026-09-29

  The product page presents Guardrails as an AWS-run service with no source repository or downloadable build: "The ApplyGuardrail API allows you to use the configurable safeguards offered by Bedrock Guardrails with any foundation model whether hosted on Amazon Bedrock or self-hosted models" - the guardrail itself is called at AWS. Searched the page for open source, license, on-prem and download: none.

• https://aws.amazon.com/service-terms/ recorded 2026-09-29

  The AWS Service Terms page states that "The Service Terms below govern your use of the Services", so use of Bedrock Guardrails is governed by AWS service terms rather than a published software license.

• https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html recorded 2026-09-29

  The user guide describes guardrails as resources created and versioned in an AWS account: "Guardrails can be used directly with FMs during the inference API invocation by specifying the guardrail ID and the version."

AWS publishes no customer count or request volume for Bedrock Guardrails, only a short list of named customers on its product page. Its reach is inferred from those named adopters and from its place inside Amazon Bedrock rather than from any published figure.

• https://aws.amazon.com/bedrock/guardrails/ recorded 2026-09-29

  The product page names six customers and gives no customer count or request volume: "Companies like Chime Financial, KONE, Panorama, Strava, Remitly, and PwC trust Bedrock Guardrails for their AI applications."

Bedrock Guardrails combines six configurable policies in one managed service, including hallucination checks through contextual grounding and Automated Reasoning, and its Standard tier extends content filters and denied topics to code. The checks run only inside AWS, whichever model they screen.

• https://aws.amazon.com/bedrock/guardrails/ recorded 2026-09-29

  The product page lists "six safeguard policies that you can configure" covering "content moderation (content and word filters), prompt attack detection, topic classification (denied topics)", PII redaction and "hallucination detection (contextual grounding and Automated Reasoning checks)", with filters for "harmful text and image content".

• https://docs.aws.amazon.com/bedrock/latest/userguide/guardrails.html recorded 2026-09-29

  The user guide describes the Standard tier extending content filters and denied topics to code: "detection of undesirable content is extended to protection against harmful content introduced within code elements including comments, variable and function names, and string literals."