> ## Documentation Index
> Fetch the complete documentation index at: https://docs.beyondguard.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# BeyondGuard: Unified AI Security for the Enterprise

> BeyondGuard secures every layer of your AI stack with specialized guards — from prompts and agents to tools and model outputs — all in one control plane.

Welcome to BeyondGuard, the enterprise AI security platform purpose-built for the age of agentic AI. BeyondGuard deploys specialized security guards across every layer of your AI stack — protecting prompts, agents, MCP/tool integrations, data pipelines, and model outputs — all managed through a single unified control plane. Whether you are running a customer-facing LLM, an autonomous agent fleet, or a RAG-powered knowledge system, BeyondGuard gives you the visibility, policy controls, and enforcement mechanisms to operate AI confidently in regulated, high-stakes environments.

## Explore the Documentation

<CardGroup cols={2}>
  <Card title="Introduction" icon="book-open" href="introduction">
    Understand why AI demands a new security approach and how BeyondGuard closes the coverage gap.
  </Card>

  <Card title="Quickstart" icon="rocket" href="quickstart">
    Connect BeyondGuard to your AI stack and deploy your first guard in minutes.
  </Card>

  <Card title="Core Concepts: Guards" icon="shield-halved" href="concepts/guards">
    Learn how BeyondGuard's five specialized guards map to the threats they prevent.
  </Card>

  <Card title="Guards Overview" icon="shield-check" href="guards/prompt-guard">
    Dive into the Prompt Guard — the first line of defense for every LLM interaction.
  </Card>
</CardGroup>

## Get Up and Running

Follow these four steps to move from zero AI-specific security controls to active threat enforcement.

<Steps>
  <Step title="Connect Your AI Stack">
    Integrate BeyondGuard into your existing AI pipeline using the SDK or REST API. BeyondGuard sits independently between your application and your models, requiring no changes to model configuration or hosting.
  </Step>

  <Step title="Configure Guards">
    Select which of the five guards — Prompt Guard, Agent Guard, MCP Guard, Data Guard, and Output Guard — apply to your deployment. Each guard targets a distinct layer of your AI value chain and can be enabled independently.
  </Step>

  <Step title="Set Policies">
    Define what BeyondGuard should do when a threat is detected: block, flag, redact, or alert. Policies are configured in the Control Plane and applied uniformly across all guarded projects without touching your model code.
  </Step>

  <Step title="Monitor and Enforce">
    Navigate to the Threat Events dashboard to review intercepted threats, audit decision logs, and track your security maturity level — from L0 (Unmanaged) through to L5 (Continuous Governance).
  </Step>
</Steps>

<Tip>
  Start with **Prompt Guard** to establish baseline visibility into your LLM interactions before layering in Agent Guard, MCP Guard, and the rest. Even L1 (Visibility) is a significant step forward from zero AI-specific controls.
</Tip>