> For the complete documentation index, see [llms.txt](https://help.genesis.autify.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://help.genesis.autify.com/settings/connection-settings/ai-provider-settings.md). # AI provider settings Autify Genesis offers two modes for managing AI providers: **Managed AI Providers** and **BYOK (Bring Your Own Key)**. Each organization selects one of the two modes when signing up. In BYOK mode, you can register API keys from your organization's AI providers in Autify Genesis and use them for AI features such as chat generation and knowledge base search. {% hint style="info" %} To switch modes after signing up, contact your Autify representative or customer support. {% endhint %} ## Prerequisites * Only the **organization Owner** can create, edit, or delete AI provider credentials. Admins and Members can only view the settings. * To newly enable **BYOK mode**, first contact . After enabling, you can manage providers by following the steps in this article. ## Managed AI providers ### Provider restrictions Manage which AI providers are available in the organization. Use the toggle on each provider to enable or disable it individually. When a provider is disabled, its models are removed from workflows and chat.
Screenshot of provider restrictions

Provider restrictions

## BYOK (Bring Your Own Key) mode ### Configuring a chat provider A chat provider is the AI provider used for chat generation. To use AI features, you must configure and enable at least one chat provider. 1. In the **Chat Setup** section, click **Configure** for the provider you want to use (OpenAI, Azure OpenAI, or Google Gemini). 2. When the setup dialog opens, enter the following information. | Field | Description | | --------------------------- | -------------------------------------- | | **API Key** | API key issued by the provider. | | **Display name (optional)** | Optional name for management purposes. | {% hint style="info" %} For **Azure OpenAI**, enter a **Base URL** in addition to the **API Key** (for example: `https://my-resource.openai.azure.com/openai/v1`). The API version is included in the base URL path and defaults to `v1`. For **Azure OpenAI** chat setup, after you click **Test Connection**, Genesis detects the deployments available in your Azure resource and shows them before you save. If no deployments are detected, you cannot save yet. Check the resource and region, or add **Deployment overrides (optional)** and test again. If your Azure deployment name differs from the model ID, specify `model=deployment` pairs in **Deployment overrides (optional)**, separated by commas (for example: `gpt-5.2=my-gpt52-deploy, gpt-5.4=my-gpt54-deploy`). Models without an override use the model ID as the deployment name. Genesis also uses the models available to your organization for background AI calls such as automatic chat title generation, memory updates, and context compaction in long workflows. {% endhint %} 3. Click **Test Connection**. * For **OpenAI** and **Google Gemini**, confirm that **Connection successful** appears. * For **Azure OpenAI** chat setup, confirm that the detected deployments are listed. {% hint style="warning" %} You cannot click **Save** until the connection test succeeds. Run the test first. {% endhint %} 4. After the test succeeds, click **Save**. 5. Turn on the toggle that appears on the provider card to enable the provider. You can register multiple chat providers. Each provider's toggle can be turned on or off independently. ### Refreshing Azure OpenAI deployments If you add or remove Azure deployments later, you can refresh the saved list without re-entering the API key. 1. On the configured **Azure OpenAI** chat provider card, click **Refresh deployments**. 2. Review the detected deployments in the preview dialog. 3. Click **Apply** to replace the saved deployment list. ### Configuring an embedding provider An embedding provider is used for knowledge base indexing and search. This is required to use knowledge base features. 1. In the **Embedding Setup** section, click **Configure** for the provider you want to use. 2. As with chat providers, enter the **API Key**. For Azure OpenAI, in addition to the **Base URL**, enter the **Deployment name**. Specify the Azure deployment name for the `text-embedding-3-small` model (1536 dimensions) in **Deployment name**. 3. Run **Test Connection**. After the test succeeds, click **Save**. 4. Select the registered provider. Genesis switches to it as soon as you click it and starts rebuilding existing knowledge bases. {% hint style="warning" %} Switching embedding providers rebuilds every knowledge base. This may take several minutes. {% endhint %} Only one embedding provider can be active at a time. ### Other AI Providers If **Other AI Providers** is available in your organization, the same settings page also lets you connect any OpenAI-compatible endpoint. These models can be used in chat and workflows. 1. In **Other AI Providers**, click **Add provider**. 2. Enter a name, the OpenAI-compatible **Base URL**, and an **API Key**. 3. Click **Test Connection**. Genesis discovers the models exposed by that endpoint. 4. Review the detected models, then click **Save**. 5. Turn on the provider to make its models available in chat and workflows. If the endpoint's available models change later, click **Refresh models** on the provider card to update the saved list. Use **Edit** to rename the provider or update its credentials, and use **Remove** to delete it. ### Editing provider credentials To change the API key or display name of a registered provider, update them with the steps below. 1. Click **Edit** on the provider card you want to change. 2. Enter the new credentials and run **Test Connection**. 3. After the test succeeds, click **Save**. ### Revoking provider credentials 1. Click **Revoke** on the provider card you want to revoke. 2. When the confirmation dialog appears, review the details and click **Revoke**. {% hint style="danger" %} Revoking credentials immediately stops every feature that uses them. There is no automatic fallback to platform-managed keys. {% endhint %} ### Troubleshooting #### Chat provider is not configured When no chat provider is enabled, a red warning message appears on the settings page. When you try to open any page other than the settings page, a setup dialog appears and blocks the operation. | Situation | Cause | Resolution | | --------------------------------------------------------------------------------------------- | ---------------------------------------------------- | -------------------------------------------------------------- | | The warning "At least one chat provider must be configured for AI features to work." appears. | No chat provider is configured, or all are disabled. | Configure at least one chat provider and turn on its toggle. | | The setup dialog appears and operations are blocked (Owner). | In BYOK mode, no chat provider is configured. | Click **Go to AI provider settings** and configure a provider. | | The setup dialog appears and operations are blocked (Admin / Member). | No chat provider is configured. | Ask the organization Owner to configure AI providers. | #### Connection test fails | Situation | Cause | Resolution | | ----------------------------------------------------------------- | ---------------------------------------------------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------- | | "Connection test failed" appears. | The API key is invalid or has expired. | Reissue a valid API key from the provider console and enter it again. | | Connection test fails for Azure OpenAI. | The base URL (resource and API version part) or the deployment settings are incorrect. | Verify the resource name, region, and deployment settings in the Azure portal, then re-enter the base URL and any required deployment overrides. | | No deployments are listed after the Azure OpenAI connection test. | Genesis could not find a supported chat deployment with the current resource or deployment settings. | Verify the resource and region. If deployment names differ from model IDs, add **Deployment overrides (optional)** and run **Test Connection** again. | Connection tests can also fail because of network settings (proxy, firewall), IP restrictions, or temporary provider outages. If the error message includes details from the provider, review those as well. #### No embedding provider is selected If registered embedding providers exist but none is selected, a yellow warning message appears: "No embedding provider selected. Knowledge base features will not work." Select one of the providers to activate it. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://help.genesis.autify.com/settings/connection-settings/ai-provider-settings.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.