Provision your Azure Bot Service agent manually

Now that you have an agent and you have tested your agent locally, you can deploy it to Azure. To run an agent on Azure, you must:

1. Provision an Azure Bot resource and configure authentication details for the resource.
2. Configure your Agents SDK agent code with details about your Azure Bot resource and authentication type.
3. Deploy your SDK agent to Azure as a web app to an Azure App Service app or container.

This article directs you to details on how to provision the Azure Bot resource and authentication resources. These instructions apply to all Azure Bot resources, regardless of which language you're using for your Agents SDK agent.

Provision Azure resources based on authentication type

The Azure AI Bot Service supports multiple authentication types. The details of how to configure each authentication type in your Azure Bot resource and in your agent code are covered in the following articles. Consider the security implications of each authentication type when making your choice.

• Provision agent resources in Azure Bot Service using client secret: For local testing (if your tenant allows secrets), use client secret and a devtunnel.
• Provision agent resources in Azure Bot Service using User-Assigned Managed Identity: For better security, use User-Assigned Managed Identity.
• Provision agent resources in Azure Bot Service using federated credentials: If you're creating a Teams agent that requires single sign-on (SSO), select federated credentials.
• Add user authorization using federated identity credential

Next steps

• Configure authentication in a .NET agent
• Configure your .NET Agent to use OAuth
• Configure authentication in a JavaScript agent
• Configure authentication in a Python agent
• Deploy your agent to Azure