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

# Quickstart

> Go from your invitation to an AI system with live guardrails, approved and ready for traffic.

By the end of this guide you will have one AI system registered, assessed, protected by live guardrails and approved to take traffic. It takes about half an hour, most of it in the assessment conversation. You need the Owner, Admin or Governance Lead role.

**1. Accept your invitation**

Open the link in the email **You are invited to join your organization on Altrum**. Set a password of at least 12 characters, or choose **Continue with Google** with the invited address. The link works once, for 7 days.

You land on [Home](https://docs.altrum.ai/home.md), which lists the setup steps until your first AI system or request arrives.

**2. Register an AI system**

Open **AI systems** and choose **Register AI system**. Give it a name, a description and a purpose, pick its owner, and answer the seven risk questions. The last step shows the rating Altrum suggests. Register it, and it starts in **Draft**.

**3. Run the assessment**

On the AI system's page, choose **Start assessment**, then **Start the assessment**. The assistant asks about the AI system one question at a time. Answer in your own words, and attach any documents that describe it. **Save and exit** keeps your place.

When the assistant has what it needs, it asks whether you are ready. Choose **Yes, submit my answers**.

**4. Review the results**

The results list the risks found, the controls that answer them, and the guardrails in scope. On the **Guardrails** tab, a guardrail that still needs choices from you is marked **Needs your action**.

**5. Set up the guardrails**

Choose **Configure guardrails**. If the banner offers suggestions from your assessment, choose **View suggestions**, then **Add them to my draft**. Finish any guardrail marked **Needs your setup**, or switch it off, and choose **Save draft**.

**6. Test the draft**

Choose **Test guardrails** and type a message that should be caught. Pick whether it is something a person sends or something the model answers. The result says whether it was blocked, masked or logged, and by which guardrail.

**7. Complete the assessment**

Choose **Complete Assessment** and confirm. Your draft becomes the live version, and the **Guardrails** tab shows **Live: version 1**.

**8. Get it approved and take it live**

On the AI system's page, choose **Request approval**. An owner or a Governance Lead approves it once its go-live checks pass. For a High or Very high rating, the approver must be someone other than the person who asked, if your organization has two or more people who can approve. See [Lifecycle](https://docs.altrum.ai/ai-systems/lifecycle.md).

Once it is approved, choose **Go live**. Applications with the AI system's keys can now reach models through the AI Gateway.

## Next

* Give the AI system a key and send its first request: [AI Gateway quickstart](https://docs.altrum.ai/ai-gateway/quickstart.md).
* Invite your colleagues and give each a role: [Members and roles](https://docs.altrum.ai/members-and-roles.md).
* Learn what each guardrail checks: [Guardrail types](https://docs.altrum.ai/guardrails/types.md).


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