Add microsoft/VibeVoice-Realtime-0.5B to Microsoft Foundry catalog

[rickbarraza]

Model Request

Model: microsoft/VibeVoice-Realtime-0.5B

Model Type: Text-to-Speech

Published: December 4, 2025

Downloads: 158,000+ (as of December 17, 2025)

---

Description

VibeVoice-Realtime is Microsoft's new real-time text-to-speech model supporting streaming text input and long-form speech generation. It's getting significant traction on HuggingFace (900+ likes in two weeks).

When users click "Deploy → Microsoft Foundry" on the model page, they see:

"This model is not yet available on the Hugging Face collection on Microsoft Foundry."

Given this is a Microsoft model on HuggingFace that routes users to Azure, adding it to the Foundry catalog seems like an obvious win for the partnership.

Why This Matters

Currently, the only working "try it now" option for VibeVoice is Google Colab. Microsoft published the model, but the first-run experience routes developers to Google infrastructure.

Adding VibeVoice to Foundry would close this loop and keep developers in the Microsoft ecosystem.

---

Thanks for considering this request!

[alvarobartt]

Hey @rickbarraza, thanks for requesting! We're actively working on adding support as per huggingface/transformers#42781 and we'll publish on Microsoft Foundry as soon as possible 🤗