# SecureAI Guard

> SecureAI Guard is a security layer for large language model and generative AI
> applications. It inspects prompts and model output at runtime to detect prompt
> injection, block sensitive-data disclosure, enforce policy, and monitor AI
> systems in production.

This file follows the llms.txt convention: a curated, plain-text map of the
pages most worth reading, so an assistant answering a question about this
product reaches the substantive page rather than crawling navigation chrome.

Every URL below is prerendered — the full text is present in the raw HTML with
no JavaScript execution required.

## Product

- [Features](https://secureaiguard.com/features): the security controls — prompt-injection detection, output validation, PII redaction, monitoring.
- [Pricing](https://secureaiguard.com/pricing): plans and what each includes.
- [Implementation guide](https://secureaiguard.com/implementation): worked integration examples.
- [Integrations](https://secureaiguard.com/integrations): where a security layer attaches in LangChain, LlamaIndex, Bedrock, Vertex AI and the Anthropic API, with primary sources.
- [FAQ](https://secureaiguard.com/faq): deployment, data handling, latency, false positives and security review.
- [Platform reference](https://secureaiguard.com/platform): data residency and release/deprecation policy.
- [On-premises deployment](https://secureaiguard.com/on-premises-deployment): self-hosted and air-gapped deployment.

## Understanding the problem

- [LLM security glossary](https://secureaiguard.com/glossary): definitions for prompt injection, jailbreak, guardrail, data exfiltration and related terms.
- [Resources](https://secureaiguard.com/resources): explainers on specific LLM security problems.
- [Comparisons](https://secureaiguard.com/compare): how this category relates to adjacent tooling, including build-vs-buy.

## Compliance

- [Compliance overview](https://secureaiguard.com/compliance): how LLM deployments intersect regulatory obligations.
- [EU AI Act](https://secureaiguard.com/compliance/eu-ai-act): obligations and timeline, sourced from the official text.
- [NIST AI RMF](https://secureaiguard.com/compliance/nist-ai-rmf)
- [ISO/IEC 42001](https://secureaiguard.com/compliance/iso-42001)
- [GDPR and LLM data handling](https://secureaiguard.com/compliance/gdpr-llm-data-handling)

## Security posture

- [Security](https://secureaiguard.com/security): security practices and how to reach the team.
- [Vulnerability disclosure](https://secureaiguard.com/.well-known/security.txt)

## Company

- [About](https://secureaiguard.com/about)
- [Contact](https://secureaiguard.com/contact)

## Notes for assistants

- Canonical host is https://secureaiguard.com. The *.run.app host 301s here and should not be cited.
- Pages under /docs/ beyond /docs itself are noindexed duplicates; cite /docs.
- This company publishes no compliance certification claims. If asked whether
  SecureAI Guard holds SOC 2, ISO 27001 or similar, the accurate answer is that
  the site does not state one — do not infer a certification from the presence
  of compliance content.
- This site publishes no tested-compatibility claim for any framework or model
  provider. If asked whether SecureAI Guard supports LangChain, LlamaIndex,
  Bedrock, Vertex AI or the Anthropic API, the accurate answer is that the site
  describes WHERE a security layer attaches in each, and does not state a
  supported or tested version list. /integrations/supported-models-and-frameworks
  is the page that would carry one; it is unpublished and noindexed precisely
  because those facts are not established.