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

# Register and manage models

> Register the AI models you want to vet and read a model's card.

Your model library is the set of models SeekrGuard can work with. Register models from OpenAI or Seekr, then use them in examinations, comparisons, chat, and risk scoring. Each registered model gets a card showing its details and its risk under a profile you choose.

## Register a model

<Steps>
  <Step title="Add Model">
    On the **Models** page, click **Add Model**.
  </Step>

  <Step title="Select Provider">
    Choose **OpenAI** or **Seekr**. The field is required.
  </Step>

  <Step title="Available Models">
    Check the models you want. **Select All** checks every model the provider offers and shows a running count, such as 158 of 158. Expanding a model reveals **Description (Optional)**, where you can add notes for your team.
  </Step>

  <Step title="Add Models">
    Click the add button. Its label counts your selection, such as **Add 1 Model**, and stays disabled until you check at least one. The new models appear in your library.
  </Step>
</Steps>

## Manage your models

The **Models** list shows one row per model, with **Name**, **Provider**, **Type**, **Parameters**, **Created**, **Status**, and **Actions**.

**Type** is foundational or fine-tuned. **Status** is active when the model is ready to use, queued while it is being prepared, or inactive when it is unavailable.

Above the table, filter by **Name**, **Provider**, **Type**, or **Status**, reorder the list with **Sort By**, and limit it to your own models with **Only mine**.

Each row's **Actions** column offers:

* **View Model Card** – open the model's card.
* **Delete Model** – available only for models you added.

## Read a model's card

**View Model Card** opens a model's card, which gathers its details, its risk under a profile you choose, and tabs for analyzing the result.

* The header shows the model's name with chips for its type and provider, its description, and its version and release date when those are available.
* **Chat with Model** opens the [Chat](/guard/app/chat) page with the model selected. **Deploy Model** is not enabled.
* The scorecard rates the model against the profile you pick from the **Risk Profile** list. It shows a composite out of 100 and a rating of strong, acceptable, marginal, or critical, weighted across the profile's categories.
* **Risk Performance** breaks the category scores down by dataset and evaluator, **Compare** sets the model against peer models, and **Explore** pivots the results by category, dataset, evaluator, or model.

<Note>
  This card's composite measures quality, where higher is better and 80 or above reads strong. The risk score on the risk pages measures the same evaluations in reverse, where higher means more risk. See [how a risk score is calculated](/guard/app/risk-profiles#understand-how-a-risk-score-is-calculated).
</Note>