> For the complete documentation index, see [llms.txt](https://docs.aisuru.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.aisuru.com/en/generative-ai/advanced/create-and-manage-model-configurations.md).

# Create and manage model configurations

AIsuru lets you create and manage access configurations for the language models (LLMs) that your Agents can use on the platform. This section will guide you through the process of creating and managing your custom model configurations.

### Accessing the models section

To manage your language models:

1. Log in to the platform;
2. Select **Models** from the sidebar;
3. You'll see a list of the configurations you've set up.

### Managing existing configurations

On the main page, you can:

* Filter configurations by purpose (Q\&A, Import/Export, Deep Thinking);
* View the details of each configuration;
* Edit existing configurations;
* Delete configurations you no longer need.

### Creating a new configuration

To configure a new model:

1. Click the **New configuration** button;
2. Fill in the required fields:
   1. Basic information:
      * **Name**: assign a unique, identifying name to the configuration; no spaces allowed, but you can use ".", "-", "&", and "@";
      * **Description**: provide a brief description of the configuration;
   2. Provider configuration:
      1. **Provider**: select the model provider (OpenAI, Anthropic, Mistral, Azure OpenAI, AWS Anthropic, Vertex Anthropic, Custom for OpenAI-compatible models);
      2. **Endpoint**:  for OpenAI, Anthropic and Mistral you can click on the "DEFAULT" button to the right of the text field otherwise, you have to enter your custom endpoint;
      3. **API Key**: enter your own API key for authentication with the provider (not mandatory for Custom provider); some providers do not require an API key, instead they require other information (f.i., client secret, secret access key, etc.);
      4. **Model**: specify the name of the model to use; some providers also require the model ID.
3. Select the purposes for which you want to enable the model:
   * Q\&A and [Board of Experts](/en/extensions/expert-group.md): for conversations with users;
   * Import/Export: for [document import](/en/agent-training/document-import.md) that creates [contents](/en/agent-training/contents.md);
   * [Deep Thought](/en/extensions/deep-thinking.md): for memory of past interactions and coordination between experts
4. Click "**Confirm**" at the bottom right to save the configuration.

#### Visibility settings (admins only)

If you have admin permissions, you'll also see a **"Visibility settings"** section when creating or editing a configuration, with the following options:

* **Visible to tenant users**: make this configuration visible and usable by tenant users;
* **Use as default**: make this configuration the default for newly created AI Agents, but only for the categories you've selected. For example:
  * If you've enabled only Q\&A, this model will be the default for user conversations;
  * If you've enabled all three categories (Q\&A and Expert Group, Import/Export, Deep Thinking), this model will be the default for all three.

Even if you set a model as default, the Agent creator can still select a different model from those available on the tenant when creating and configuring the Agent.

{% hint style="info" %}
If you have a [separate AIsuru tenant](/en/paas.md), keep in mind that **a model that isn't visible to other users is still always visible and usable by other tenant admins**.
{% endhint %}

#### Duplicating an existing model

You can **duplicate** an existing model to quickly create a new one based on the same configuration.

**To duplicate a model**:

1. Go to the **Models** tab in your Agent's settings;
2. Find the model you want to duplicate in the list;
3. Click the **Duplicate** icon (the middle one between "Edit" and "Delete");
4. A pre-filled form will open with all the fields from the original model.

**In the duplication form, you can**:

* Change the **name** of the configuration;
* Modify any field before saving;
* Click **Confirm** to create the new duplicated model.

> **💡** Duplication is useful when you want to create variations of the same model with small changes (e.g. same provider but different location, or same configuration but different model). Please, note that visibility settings will be duplicated as well. You can change them before saving the new configuration.

### Bulk-apply a configuration

If you have admin permissions and want to apply the same model configuration to all Agents on the tenant, you can use the **"Apply to all"** feature:

1. **Prerequisites**:
   * The "**Use as default**" flag must be enabled;
   * The "**Visible to tenant users**" flag must be enabled;
2. **How to use it**:
   * After configuring all the model parameters, click the **"Apply to all"** button;
   * Select the scope: Q\&A and Board of Experts, Import/Export, Deep Thought or Enable for all purposes.

The configuration will be automatically applied to all existing Agents on the tenant.

{% hint style="danger" %}
This action **will overwrite the existing model configurations for all Agents on the tenant**. Use this feature to standardize the use of a specific model across all Agents on the platform.
{% endhint %}

### Tips

* Use clear names and descriptions to make management easier;
* Double-check all data before saving a configuration;
* Only enable the purposes that are relevant to each model;
* Review configurations periodically to keep them up to date.

Keeping your configurations well-organized helps optimize your Agent's performance, ensuring more accurate and relevant responses for your specific needs.