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

# Compliance

> Measure your agents against the frameworks you work to, and share the evidence with your auditors.

Show your auditors where your agents stand, with evidence drawn from what they actually did. Compliance measures your agents against the frameworks you choose, control by control.

For example, choose the EU AI Act. Each control shows whether it is **Verified**, **Failing** or **Monitoring**, the policies in force for it, and the turns that are its evidence. **Share with auditor** gives your auditor a read-only link to the same evidence.

> [!NOTE]
>
> Compliance is in beta and off by default. It is turned on per organization by your Xenovia contact, or by the administrator of a self-hosted deployment; there is no switch for it in the console. Once it's on, **Compliance** appears in the sidebar with a **Beta** badge.

## Choose your frameworks

The first time, an admin chooses **Choose frameworks**: EU AI Act, NIST AI RMF, OWASP LLM, SOC 2, GDPR, HIPAA or PCI DSS. Xenovia suggests some from the data your agents handle. Change them later with **Settings** on the Compliance page.

## The Compliance page

The header shows how many of the framework's controls are verified and how many are failing, and when they were last checked. Pick a **Framework**, then one of four tabs:

| Tab | What it shows |
| - | - |
| **Controls** | Each control's **Status**, **Coverage** and **Evidence** (turns and errors) |
| **Checks** | The automated checks, run every 15 minutes, and their latest result |
| **Agents** | Each agent's owner, risk tier (EU AI Act: Minimal, Limited or High), the data it handles (personal, health or card data), and its config. **Classify** sets them |
| **Baselines** | Policies that apply across agents, in **Enforce**, **Monitor** or **Off**, and exemptions for an agent, up to 90 days |

A control's status comes from its automated checks, from what the runtimes did with the policies mapped to it over live traffic, and from the agents' profiles. It is one of **Verified**, **Failing**, **Monitoring**, **Configured**, **Not checked yet** or **Not covered**. When telemetry is unavailable, statuses that rest on what the runtimes did read **Not checked yet**.

Open a control to see its automated checks, the **Policies in force**, **Prebuilt policies** you can add with **Use template**, exemptions, and failures in the last 90 days. The **To do** list offers **Cover** for gaps: it creates baselines that monitor and block nothing until you enforce them.

![Compliance page for the EU AI Act with a to-do list and controls marked Verified, Monitoring, Failing and Not covered](https://mintcdn.com/xenovia/RO5UF1Tk8L9B7yPt/images/screens/compliance-controls.png?fit=max\&auto=format\&n=RO5UF1Tk8L9B7yPt\&q=85\&s=624838709126f13c309c31b9c9d1669a)

The Compliance page, Controls tab

Admins can **Run checks now**. When Nova is on, **Ask Nova** on a control asks where you stand on it and how to close the gap.

## Share the evidence

* **Download evidence pack** downloads the evidence for the framework, signed with the audit key when one is set up.
* **Share with auditor** creates a read-only link to the live evidence pack. Say **Who it is for** and when the link expires (7 to 90 days). Copy it at once: it is not stored and can't be shown again. Every time the link is opened, it is recorded in the [audit log](https://docs.xenovia.io/organization/audit-logs.md). **Revoke** cuts access at once.

The auditor's page shows the controls today, failures in the last 90 days, the agents in scope, the baselines, the audit trail's integrity and whether the pack is signed.

## Alerts

With Compliance on, an [alert rule](https://docs.xenovia.io/observability/alerts.md) can fire on **Compliance check starts failing**.

## What next?

> [card] [Audit logs](https://docs.xenovia.io/organization/audit-logs.md)
>
> The signed record behind the evidence.

> [card] [Policies](https://docs.xenovia.io/control/policies.md)
>
> Write the rules a control needs.

> [card] [Alerts](https://docs.xenovia.io/observability/alerts.md)
>
> Be told when a check starts failing.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.