Gnani.ai Launches Vachana Speech-to-Text Model Under IndiaAI Mission

Gnani.ai, India’s conversational AI company, announced the launch of Vachana speech-to-text (STT), an Indic STT model trained on over one million hours of real-world voice data, as part of its selection under the Indian government’s IndiaAI Mission. 

The model is designed to support enterprise speech recognition across multiple Indian languages and sectors, the company said.

“Speech recognition in India is not a localisation problem. It is a foundational systems problem,” said Ganesh Gopalan, co-founder and chief executive officer of Gnani.ai. “Vachana STT is built as core infrastructure, trained on how India actually speaks, and designed to operate across channels.”

The Bengaluru-based company said Vachana STT forms the first release in its upcoming VoiceOS stack and is available immediately via API for enterprise customers. According to the company, early adopters will receive one lakh free minutes of usage.

Gnani.ai said in a statement that the model has been trained on proprietary multilingual datasets spanning more than 1,056 domains. It supports real-time and batch transcription and is already deployed across banking, telecom and customer support operations, collectively processing about 10 million calls per day with a P95 latency of 200 milliseconds.

In internal and public dataset evaluations, Vachana STT recorded 30-40% lower word error rates for low-resource Indian languages and 10-20% lower error rates for the eight most-used languages in India, the company said. 

The evaluations covered languages including Hindi, Tamil, Telugu, Kannada, Bengali and Marathi.

The model is built to handle compressed audio, variable network conditions and high concurrency, making it suitable for compliance monitoring, analytics and voice-driven workflows, Gnani.ai said. 

The company added its selection under the IndiaAI Mission reflects a focus on building sovereign foundational AI infrastructure rather than application-layer tools.