Kiriku-Pulaar-TTS
Pulaar speech synthesis; restricted access
Official site consultedAI: models
Last verified: 1 October 2026
Single-speaker Pulaar text-to-speech model of VITS type, via the Coqui TTS framework, trained directly on graphemes without a phonemizer. It outputs 22,050 Hz audio with a checkpoint of about 952 MB. Gated repository.
Services
- Pulaar text-to-speech
Access requirements
Gated repository: login and sharing of contact information required; license not specified.
Sources
- Kiriku-Pulaar-TTS (new tab), verified on 1 October 2026Official site consulted