MAI-Voice-1
by Microsoft
· Mai family · released Oct 7, 2025
Open weights
Active
Microsoft's open-source neural text-to-speech and voice cloning model with strong speaker similarity. Released on Hugging Face and integrated into Azure AI Speech.
This model has no first-party API. Open-weight release: self-host from published weights.
Specifications
| Model type | Audio gen |
|---|---|
| Architecture | Unknown |
| Parameters | Not disclosed |
| Context window | Not disclosed tokens |
| Max output | Not disclosed tokens |
| Knowledge cutoff | Not disclosed |
| License | MIT License |
| Modalities | Text input Audio input Audio output |
| Tags | - |
| End of life | Not disclosed |
| Last verified | Not disclosed |