Share via

Azure text-to-speech-avatar not yet available for deployment

Janneke van de Loo 5 Reputation points
2026-05-08T09:12:25.8233333+00:00

I want to try out the Azure text-to-speech-avatar model, but when I select the model for deployment in the AI Foundry, I see a message "model is not yet available for deployment". This happened both in the West Europe region and in the Sweden region. When will the Azure text-to-speech-avatar model be available for deployment in those regions? Are there other regions where it is available for deployment? And until then, is there another way to try out the avatar model?

Azure Speech in Foundry Tools

2 answers

Sort by: Most helpful
Most helpful Newest Oldest
  1. Anshika Varshney 11,675 Reputation points Microsoft External Staff Moderator
    2026-05-08T10:38:11.0033333+00:00

    Hi Janneke van de Loo,

    This behavior is expected currently and not an issue with your setup.

    The main point to understand is that Text-to-Speech Avatar is not deployed like Azure OpenAI models in AI Foundry. Even if you see it in the catalog, it is not yet available as a deployable model, which is why you get the message “model is not yet available for deployment.

    What is actually happening

    • Text-to-speech avatar is a Speech service feature, not a standard deployable model in Foundry.
    • In some regions like West Europe or Sweden Central, the feature exists, but deployment through Foundry is not supported yet.
    • There is no public timeline yet for when it will be available as a deployable model.

    How you can use it (working approach)

    Instead of trying to deploy it from Foundry, you can use it through Azure Speech Service APIs or tools:

    1. Create a Speech resource in a supported region
    2. Use Speech Studio or SDK/API to generate avatar videos
    3. Try:
    • Real-time avatar (interactive)
      • Batch avatar (video generation)

    You can follow the official quickstart here:

    Supported regions

    Make sure your Speech resource is in a region where avatar is supported (for example):

    • East US 2
    • North Europe
    • South Central US
    • Southeast Asia
    • Sweden Central
    • West Europe
    • West US 2

    You can also check the official region list here:

    Common things to double check

    • You are using a Speech resource (not only Foundry model deployment)
    • Region supports avatar feature
    • Using standard S0 pricing tier (needed for avatar features) [learn.microsoft.com]
    • Trying via Speech Studio or SDK instead of model deployment

    In short:

    • The error is not a bug
    • Avatar cannot be deployed as a model in Foundry yet
    • You can still use it via Speech service APIs or Speech Studio

    Do let me know if you have any further queries.
    Thankyou!

    Was this answer helpful?

    0 comments

  2. AI answer

    2026-05-08T09:12:39.74+00:00

    Azure text-to-speech avatar is currently only available in specific Azure Speech regions and is not exposed as a deployable model in Azure AI Foundry in the same way as Azure OpenAI models.

    1. Region availability Text-to-speech avatar is supported as a Speech service feature (not an Azure OpenAI model) in these regions:
    • eastus2
    • northeurope
    • southcentralus
    • southeastasia
    • swedencentral
    • westeurope
    • westus2

    In these regions, the following capabilities are available:

    • Real-time avatar
    • Batch avatar
    • Custom avatar
    • Custom avatar training (only in southeastasia, westeurope, and westus2)

    If AI Foundry shows “model is not yet available for deployment” for an avatar-related entry in West Europe or Sweden Central, that means the avatar model is not yet offered as a Foundry-deployable model in that environment, even though the underlying Speech avatar feature is available in those regions.

    1. When it will be available in Foundry The context does not provide any timeline or roadmap for when text-to-speech avatar will become deployable as a model in Azure AI Foundry in West Europe, Sweden Central, or any other region. Model and feature availability is centrally controlled and cannot be requested per subscription.
    2. How to try the avatar feature today Until a Foundry deployment option is available, the supported way to use Azure text-to-speech avatar is via the Azure Speech service APIs in one of the supported Speech regions listed above. Key points:
    • Use a Speech resource in a region that supports text-to-speech avatar (for example, westeurope or swedencentral).
    • Call the text-to-speech avatar APIs directly (real-time or batch) instead of trying to deploy a model in AI Foundry.
    • For real-time avatar with Voice Live, specify the avatar parameter and follow the WebRTC/SDP flow:
      • Start a Voice Live session.
      • Include the avatar parameter to enable synchronized avatar output.
      • Optionally provide ice_servers, or use the ICE servers returned in session.updated.
      • After ICE candidates are gathered, send: 
            {
        "type": "session.avatar.connect",
        "client_sdp": "your-client-sdp"
    }
      • The service responds with: 
            {
        "type": "session.avatar.connecting",
        "server_sdp": "your-server-sdp"
    }
      • Use server_sdp to establish the avatar media connection.

    This approach lets applications use text-to-speech avatar now, even though it is not yet available as a deployable model in AI Foundry.

    References:

    AI-generated content may be incorrect. Read our transparency notes for more information.

    Was this answer helpful?

Your answer

Answers can be marked as 'Accepted' by the question author and 'Recommended' by moderators, which helps users know the answer solved the author's problem.