Repository navigation
Maya 2 Native is Maya Research’s open-weight, streaming text-to-speech model for 11 Indic languages and English, with 20 built-in voices, cross-lingual synthesis, and zero-shot voice cloning from 3–10 seconds of reference speech.
- Download the weights and hear samples
- Read the launch article
- Installation and inference examples
- Evaluation setup, results, and limitations
The inference engine provides a Python API, an OpenAI-compatible speech endpoint, and WebSocket streaming. Audio is 24 kHz mono. Start with max_seats=1 and consult the memory table before increasing concurrency.
The public model name is Maya 2 Native. Existing package names, API identifiers, configuration filenames, and weight filenames are preserved, including maya3, Maya3TTS, and maya3_config.json.
Weights and inference code are Apache-2.0; upstream components retain their respective licenses. Clone only voices you own or have permission to use, and disclose synthetic speech.
Maya Research team: Dheemanth Reddy BS, Bharath Kumar Kakumani, Sai Charan Thippareddy, and Prahalad Boreddy. Thanks to Qwen, Kyutai, SpeechBrain, and the Hugging Face community.