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

# Build agents that govern themselves

> Intent Detection and Response (IDR) for the age of AI agents — every prompt declares an intention, every action runs through policy, every credential is brokered short-lived.

Inviolet is the **Intent Detection and Response (IDR)** platform — every prompt
declares an intention, every action runs through a layered policy, and every
credential is brokered short-lived against your existing PAM. This is the
technical documentation; for product overviews and pricing, see
[inviolet.ai](https://inviolet.ai).

## Start here

<CardGroup cols={2}>
  <Card title="Quickstart" icon="rocket" href="/quickstart">
    Zero to a governed tool call in five minutes.
  </Card>

  <Card title="Concepts" icon="book-open" href="/concepts/intentional-access">
    What Intent Detection and Response (IDR) is, why it matters, and where it sits in
    your stack.
  </Card>

  <Card title="SDKs" icon="code" href="/sdks/node">
    Node, framework wrappers, Python, and REST clients for the Inviolet gateway.
  </Card>

  <Card title="Tier comparison" icon="layer-group" href="/tiers/overview">
    Violet, Deep Violet, and Ultraviolet — what each tier unlocks.
  </Card>
</CardGroup>

## Three minutes to your first governed call

1. **Install the SDK** for your stack — [Anthropic](/sdks/anthropic),
   [OpenAI](/sdks/openai), [fetch](/sdks/fetch), [LangChain](/sdks/langchain),
   [LlamaIndex](/sdks/llamaindex), or [MCP](/sdks/mcp-client).
2. **Create an intent card** in the Gallery — name the purpose, scope the
   columns, set the response strategy.
3. **Guard your agent** with `new Inviolet({ gatewayUrl, apiKey, agentId })` —
   `mintMandate()` at the start of a session, then `guard()` every tool call.

That's it. Open the dashboard and watch your first decision flow through the
cascade in real time. Full walkthrough in the [Quickstart](/quickstart).

## Why IDR?

Traditional access control asks *"who is calling?"*. Modern agents need a
sharper question: *"under what intention?"*. IDR is the architecture that makes
"intention" a first-class signal — observable, enforceable, brokered.

The 5 Principles are the doctrine. Every API endpoint, every SDK wrapper, every
dashboard tile traces back to one of them.

## The 5 Principles, in 30 seconds

1. **Learn the intent behind every prompt.** No declared intent → no action.
2. **Test before enforcing.** Every policy can run shadow → enforce.
3. **Least access as the system learns.** Column-grain, not connection; the gray zone routes to a human.
4. **Agents never touch credentials.** Inviolet brokers a short-lived lease against your PAM.
5. **Impossible to bypass.** Every path covered and proven — SWG egress, every LLM host, bypass-tested in CI.

Each principle maps 1:1 onto one of the five levels of the
[maturity model](/concepts/maturity-model). Read the full doctrine in
[The 5 Principles](/concepts/principles).

## What this documentation covers

* **Getting Started** — concepts and installation paths
* **Guides** — step-by-step recipes for the most common workflows, plus
  day-two [operations](/operate/dashboard)
* **API Reference** — every gateway endpoint
* **SDKs** — the `Inviolet` client, framework wrappers, and the REST flow

For commercial questions, talk to the
[Inviolet sales agent](https://inviolet.ai) on the marketing site — it has full
context on tiers, pricing, and your specific compliance requirements.