The Voice AI Revolution Brewing in Bengaluru

Dheemanth Reddy and Bharath Kumar from Maya Research AI have developed an open-source, world-class voice conversational models from their SPC Bengaluru office in HSR Layout. Their 3B parameter Maya1 model is open-weight and was trained entirely using free compute credits.

Anyone can download the model, inspect it, use it, or fine-tune it. It’s not locked or proprietary like many large AI models. The model can understand speech and respond in natural sounding speech — similar to Siri, Alexa, or GPT voice chat. The model is strong enough to be compared with leading voice AI models, despite being built with minimal resources.

Core Capabilities:

• 20+ emotional styles (e.g., cheerful, calm, dramatic).
• Zero-shot voice design - clone or design new voices without needing training data.
• 3B-parameter architecture, optimized for production-ready real-time streaming.
• Apache 2.0 license - businesses can deploy and monetize with no per-usage fees.
• Supports fine-tuning to create unique brand or character voices.

It is comparable to closed tools like ElevenLabs or Murf.ai but with a major advantage: You own the deployment + pay no per-second fees.

Maya Research's other model, Veena is a text-to-speech model designed especially for Indian audiences. Its standout ability is smooth Hinglish code-switching—it can switch between Hindi and English naturally in the same sentence, the way people in India commonly speak.

Key highlights:

• Low latency speech generation (under 80 ms) — ideal for real-time applications.
• Open source (Apache 2.0 license) — developers can use it freely, even commercially.
• Trained on high-quality proprietary speech datasets.
• Uses 15,000+ spoken utterances per speaker, contributing to its clarity and realism.

This is the kind of open, developer-forward innovation that can accelerate voice AI adoption across India’s apps, media, and creator ecosystem.