How to integrate the live chat avatars we created in Azure speech service into MS teams Meetings?

Hello Prasanna Sai Kommineni,

Welcome to Microsoft Q&A and Thank you for your question.

Integrating the custom avatars you’ve created using Azure Speech Services into Microsoft Teams meetings is definitely an interesting scenario,

Make sure your avatar is fully trained and deployed in Azure Speech Services. You can create, train, and manage the model through Azure AI Foundry following the Custom Text-to-Speech Avatar guidance. Once training is complete, deploy the model to an endpoint so it’s available for real-time usage.

Teams does not currently offer a direct or native way to plug Azure Speech real-time avatars into a meeting as a camera feed. However, you still have a few practical ways to use your custom avatar within Teams meetings depending on your use case.

If your solution involves interacting with Teams meetings programmatically, Azure Communication Services (ACS) provides a way for an application to join Teams meetings. The ACS quickstart for “Join your app to a Teams meeting” is a good reference for building the meeting experience, especially for scenarios where your bot or app needs to participate in the meeting.

Once your avatar model is deployed, you can use its REST APIs to generate real-time video or audio output. That output can then be surfaced in Teams meetings in one of the following ways:

• Virtual Camera Approach: Render the avatar video stream and route it through a virtual camera driver (e.g., OBS Virtual Camera). Teams can then pick that virtual camera as the video input.
• Shared Screen/Window Approach: Host your avatar in a web or desktop app, and share that screen directly inside the meeting.
• Pre-rendered Output in Bots: If you are using a Teams bot, you can include pre-rendered avatar videos or messages, though bots cannot send a live video feed today.

Custom avatars are restricted to qualified customers and partners under Microsoft’s responsible AI policies. Please ensure your use case is aligned with these requirements.

Please refer this

1. How to create a custom text-to-speech avatar
2. Quickstart: Join your chat app to a Teams meeting

I Hope this helps. Do let me know if you have any further queries.

Thank you!