Shunya Labs Launches Vāķ, India's Largest Open-Weight Voice AI Supporting 55 Indian Languages

Shunya Labs, in partnership with Nasscom, has launched Vāķ at the India AI Impact Summit 2026, marking a significant advancement in India's linguistic AI infrastructure. The system supports 55 Indian mother tongues, enabling real-time translation across all 2,970 possible language pairs with end-to-end latency under 1.5 seconds while preserving speaker voice, tone, and emotion.

The release addresses a critical gap in India's AI ecosystem, where most speech systems support only 5-10 Indian languages, forcing reliance on foreign APIs and leaving major speech communities like Bhojpuri, Rajasthani, Chhattisgarhi, and Magahi underserved. Vāķ's open-weight architecture allows organizations to deploy the system on their own infrastructure without data leaving the country, eliminating recurring API costs and foreign dependencies.

The platform comprises three model families: Pingala ASR, ranked #1 on the Hugging Face OpenASR Leaderboard with 3.10% Word Error Rate; a neural TTS engine with zero-shot voice cloning; and the real-time translation system. Coverage spans 43 Indo-Aryan, seven Dravidian, three Sino-Tibetan languages, Santali, and Indian English, reaching over 1.17 billion native speakers and aligning with initiatives like the *IndiaAI Mission*, *Digital India*, and *BHASHINI*.