For the complete documentation index, see llms.txt. This page is also available as Markdown.

Guardrails

Available guardrails, grouped by primary category. Select a guardrail to view its API details. See the Taxonomy reference for what each category means.

Query this catalog programmatically with AnyGuardrail.list_guardrails(...) and AnyGuardrail.group_by(...) — see the AnyGuardrail reference.

Prompt Injection

Guardrail
Description

Hosted detector for direct (user prompt) and indirect (document-borne) prompt-injection and jailbreak attacks.

Binary prompt-injection classifier.

Binary safety and jailbreak classifier, scoring a prompt or prompt-response pair.

Binary prompt-injection classifiers.

Hosted API for prompt-injection, jailbreak, content-moderation, and PII detection.

Binary prompt-injection classifier.

Binary prompt-injection classifier trained to mitigate over-defense.

Encoder classifier for prompt-injection and jailbreak detection.

Binary prompt-injection classifiers.

Binary prompt-injection classifier.

Binary prompt-injection and jailbreak classifier using a chunked e5-large encoder with a trained MLP head.

Content Safety

Guardrail
Description

Hosted content-moderation and safety-detection API with configurable detection policies.

Hosted moderation of text and images across hate, sexual, self-harm, and violence categories with 0-7 severity scores.

Hosted, configurable moderation covering content filters, denied topics, PII, word filters, and contextual grounding.

Polish multi-label safety classifier.

Multilingual multi-label safety classifier scoring text across 12 harm categories including jailbreak prompts.

Schema-driven safety, toxicity, jailbreak, and refusal detector.

Policy-grounded reasoning safety classifier that judges text against a bring-your-own written policy.

Korean safety decoder models covering harmful content and legal risk.

Decoder-LLM safety classifier judging prompts and responses against the 14-category MLCommons hazard taxonomy.

Reasoning safety classifier covering a 22-category content-safety taxonomy.

Hosted moderation API flagging content across 13 harm categories with calibrated scores.

Multilingual safety-moderation judge reporting request harm, response harm, and refusal across 17 languages.

Generative safety moderation with three-level severity across 119 languages.

Token-level streaming safety moderation with span output.

Policy-conditioned safety classifier that judges a prompt against a user-supplied policy via Yes/No token logits.

Hosted text-detection moderation API running configurable Granite Guardian detectors.

One-pass safety-moderation judge reporting prompt harm, response harm, and refusal.

PII

Guardrail
Description

Span-level PII/NER detector emitting character spans and a redacted copy of the text.

Hallucination

Guardrail
Description

Token/span-level RAG hallucination detector.

Off-Topic

Guardrail
Description

Cross-encoder relevance detector that flags whether an input strays from a comparison text.

General Judge

Guardrail
Description

Policy-based LLM judge that grades text against a natural-language policy using an LLM provider.

Generalist LLM judge that scores a response against user-defined criteria and rubric on a 1-10 scale.

Dynamic guardian model evaluating conversation compliance with user-defined policies.

Local LLM judge scoring text against user-defined criteria, metrics, and rubrics.

Prompt-based LLM judge that grades text against user-supplied pass criteria and rubric, returning reasoning and highlighted phrases.

Hybrid-thinking safety and judge model covering harm, RAG groundedness, and function-calling risks via bring-your-own-criteria.

Hosted evaluation API running configurable evaluators for hallucination, toxicity, PII, prompt injection, and custom judging.

Open rubric-based LLM judge grading a response on a user-defined 1-5 rubric.

General-purpose LLM judge grading a response against a user-defined 1-5 rubric.

Last updated