Maya2 Native brings streaming TTS to Indic languages

Maya Research has released Maya2 Native, an open text-to-speech model built for streaming audio generation with a clear emphasis on Indian languages. The model supports voice cloning and multilingual synthesis, with its stated focus on the Indic language family.

Streaming is the practical centerpiece here. Rather than producing an entire clip before playback, a streaming TTS system emits audio incrementally, which keeps latency low enough for conversational and interactive applications. Paired with voice cloning, that makes the model a candidate for assistants, dubbing, and accessibility tools that need to speak naturally in regional languages.

Why it matters

High-quality speech synthesis has long skewed toward English and a handful of global languages, leaving India's many widely spoken tongues underserved. A model that treats Indic languages as a first-class target helps close that gap:

• Streaming output for low-latency, real-time use
• Voice cloning from reference audio
• Multilingual coverage anchored in Indian languages

The release is distributed on Hugging Face under a non-standard license, so teams should review the terms before building on it. As an initial entry in the Maya2 line, it will be worth watching how the model's voice quality and language coverage hold up against established open TTS systems in real deployments.