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

# Quickstart

> Sign in to ModelHub with an email code or SSO and send your first message in a few minutes.

This page gets you from a link to your first answer. It assumes an administrator has already approved your email and connected at least one model. If not, see [Rolling out ModelHub](/modelhub/admin/rollout).

<Steps>
  <Step title="Open ModelHub">
    Go to [https://modelhub.corepillar.ai](https://modelhub.corepillar.ai). If you are not signed in you land on the **CorePillar** sign-in card.

    <Tip>If an administrator sent you an invite link (it ends in `/invite/` followed by a token), open that link instead. It signs you in and adds you to the right organization in one go.</Tip>
  </Step>

  <Step title="Enter your work email">
    Type your email in the **Email** field.

    * If your organization uses single sign-on, a **Sign in with SSO** button appears after a moment. Click it and complete your company's Okta sign-in.
    * Otherwise click **Send me a code**. CorePillar emails you a 6-digit code.

    If you see "Your access is pending approval by an administrator.", your email has not been approved yet. Ask your administrator; there is nothing else to do on your side.
  </Step>

  <Step title="Enter the code">
    Type the code into **6-digit code** and click **Verify & sign in**. You land on the chat screen of your organization's first project that has a model connected.
  </Step>

  <Step title="Ask something">
    The home screen shows a greeting, the message box and four example prompts. Click an example or type your own question in **Message the assistant…**, then press **Enter** (or click **Send**).

    Leave the model picker on **Auto** for now. ModelHub picks the cheapest connected model that can answer.
  </Step>

  <Step title="Read the answer and its report card">
    The answer streams in as it is written. When it finishes, a card under the answer shows:

    * a verdict (**Right-sized**, **Overkill** or **Underpowered**) and a one-line explanation,
    * the cost of the answer and how long it took,
    * a **Details** button with the routing signals, privacy findings and token counts.
  </Step>

  <Step title="Keep going or start fresh">
    Reply in the same box to continue the conversation. Click **New chat** at the top of the sidebar to start a new one. Your conversation appears in the sidebar under **Today**.
  </Step>
</Steps>

## Try the privacy screening

Send a message containing something personal, for example "Draft a reminder email to [jane.doe@example.com](mailto:jane.doe@example.com) about Tuesday". The address in your sent message turns into a dashed chip such as `[EMAIL_4821]`. Hover it to see the original value; the model never received it. See [Privacy screening](/modelhub/chat/privacy-screening).

## Next

* [Chat overview](/modelhub/chat/overview) explains every control on the screen.
* [Models and routing](/modelhub/chat/models-and-routing) explains Auto and when to pick a model yourself.
* [AI Usage](/modelhub/usage) shows what your conversations cost.


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