|
PNN
Mumbai (Maharashtra) [India], September 24: Navana.ai, a leading sovereign voice AI company for regulated enterprises whose systems have handled more than 100 million conversations for Indian banks and lenders, today unveiled and opened public access to Bodhi TTS, a text-to-speech voice AI model built from the ground up for real-time Indian voice AI applications. It launches with more than 50 natively Indian voices across 10 Indian languages at one list price of 12 per 10,000 characters, roughly 0.72 per minute of generated speech. Bodhi TTS is sovereign by architecture, not merely by hosting. Navana.ai trains and owns the entire Bodhi speech stack, with no information routed through a third-party model provider. The model, which is live in 10 Indian languages, including Hindi, Telugu, Tamil, Marathi, Bengali, Odia, Malayalam, Kannada, English, and Gujarati, is approximately five times smaller than the nearest comparable architecture, enabling compelling pricing and inference to run inside a banks own data centre. This allows customer audio, transcripts, or personal data to remain within the enterprises security perimeter, supporting the security requirements of regulated businesses. Navana.ai is ISO 27001 and SOC 2 Type II certified. What makes Bodhi work in India is how the model is built. The model handles Indian text natively: 1,00,000 is read as one lakh rupees, and PAN numbers, phone numbers, dates, EMIs, scheme names and account numbers are spoken the way a bank would say them. Pronunciation is controlled based on sounds: a BFSI enterprise customer uploads a dictionary of their product names, branch names and customer names, and the language model driving a conversation can mark up a name inline and be obeyed exactly. Any mispronunciation flagged on a live call is fixed in minutes rather than a retraining cycle. The 50+ voices are natively Indian tuned for natural, conversational instructions, reflecting how people speak in real-world phone conversations rather than studio narration. New voiceS can also be added from a few seconds of audio without additional training. All features are provided as part of the standard product offering without any additional charges. Every voice costs the same, with no premium tier. At 12 per 10,000 characters, Bodhi TTS is significantly more cost-effective than comparable offerings. ElevenLabs lists its Multilingual and v3 models at $0.10 per 1,000 characters, about 95 per 10,000; Sarvam AI lists Bulbul v3 at 30. Announcing the official rollout of Bodhi TTS, Raoul Nanavati, Co-Founder & CEO of Navana.ai, said, India runs on voice, and most of it is not in English. Until now, a bank that wanted a voice agent in Marathi or Tamil had to choose between a global model that mispronounced everyday words and a price that only worked for a pilot. We built Bodhi from the ground up to work to represent every accent, so in a country where every region speaks a little differently, businesses get fine-grained control from day one to make their bots sound truly Indian. And we priced it for the call volumes Indian businesses actually run, with everything in the box. One price, every voice, every feature. Built for how Indians actually speak Bodhi TTS begins producing audio in under 100 milliseconds, enabling callers to hear responses almost immediately while the remainder of the sentence is being generated. Voices carry pitch, pace and vocal character from a reference clip with no emotion tagging, enabling more natural and conversational interactions. The model also offers zero-shot cloning: a few seconds of audio, added live over the API with no redeploy in all 10 languages. Reference audio is cached, so a custom voice adds no latency to a call. Everyone assumes quality in speech scales with model size. For Indian speech it does not, said Jai Nanavati, Co-Founder of Navana.ai. We built Bodhi TTS about five times smaller than the nearest comparable model. A small, fast model is one a bank can run on a mid-tier GPU it already owns, inside its own data centre. That is why we can price Bodhi the way we have, and why no customer data ever has to leave the building, he added. Bodhi models sit behind voice agents that support more than 1,000 crore in monthly loan disbursals across Navana.ais banking and lending clients. Bodhi STT, the companys speech recognition model for 10 Indian languages and more than 40 dialects, opens by invitation the same day, with benchmark results published on Hugging Face. Bodhi TTS is open to all starting today at https://platform.navana.ai, with 1,000 in free credits on signup. Banks, insurers, and lenders can apply for early access to both models; the first 100 approved teams receive 100 hours of usage credits, valid for 90 days. Since its founding, Navana.ai has serviced over 100 million voice AI minutes with leading BFSI players such as Bajaj Finserv, Protean, Ujjivan Small Finance Bank, Jana Small Finance Bank, Grihum Housing Finance, and others to automate their call center-driven business outcomes across sales, collections, compliance, and support with Voice AI. About Navana.ai Navana.ai is building Indias sovereign voice AI infrastructure for regulated industries. Founded by Raoul Nanavati and Jai Nanavati, the companys Bodhi speech models power voice AI across Indias leading BFSI companies, enabling enterprises to serve customers in their own language at pan-India scale. Navana.ai is ISO 27001 and SOC 2 Type II certified. Learn more atnavana.ai. (ADVERTORIAL DISCLAIMER: The above press release has been provided by PNN. ANI will not be responsible in any way for the content of the same.)
|
