smallest.ai, an AI startup headquartered in San Francisco, California focused on multi-modal models, introduced Lightning, text to speech (TTS) model capable of generating up to 10 seconds of audio within 100 milliseconds.
This advancement enables developers worldwide to build highly realistic voicebot applications with sub second latency, streamlining the implementation process and making it accessible at significantly lower costs.
Lightning supports English and Hindi in multiple accents currently and the team plans to add many more languages quickly.
Being priced as low as $0.02 (or approximately 1.6 Rs) per minute, Lightning provides a cost effective solution, enabling the applications to run at under 1 Rs/minute—drastically reducing expenses for voicebot builders and broadening market accessibility.
Lightning’s rapid processing and cost efficiency make it a notable alternative in the voicebot industry, where traditional TTS models often rely on streaming and web sockets, increasing server demands and complicating scalability.
This has been designed for practical and high speed integration, Lightning operates through a straightforward REST API, delivering audio in around 100 milliseconds without the server strain associated with continuous streaming. Currently supporting multiple English and Hindi accents, Smallest.ai plans to expand the model’s language range to include other Indian, European, and Asian languages in the coming months.
Smallest.ai was founded by IIT Guwahati alumni Sudarshan Kamath and Akshat Mandloi. Kamath attributes the affordability of smallest.ai to their focus on data quality and model efficiency. “Our model is much smaller than those of competitors like ElevenLabs. Despite this, we achieve high-quality speech because our data is highly refined,” he explained.
Voicebot developers with early access to Lightning have reported an 8x reduction in operational costs, accompanied by enhanced audio quality. Beyond real-time voicebot applications, Lightning is adaptable for creating audiobooks and voiceovers for social media content on platforms like Instagram and YouTube.
Non-developers can access Lightning through the Waves Speech platform, where additional features, including voice cloning and accent conversion, are also available in beta.
“When we started building, we realised that the models required for a voice bot were not mature for Indian languages. Existing models for non-English languages were nowhere close to production,” explained Kamath in an exclusive interaction with AIM.
Earlier in June, smallest.ai also launched AWAAZ allows voice cloning from short audio clips and is available at competitive rates. The model is aimed at scalable applications in regional language markets and provides enterprise-grade security and compliance.
When asked about its mission, Kamath said, “Why are 1B humans not speaking to AI voices everyday despite incredible advancements in Voice AI? This is the problem we are trying to solve.”
The post Bengaluru AI Startup smallest.ai Unveils Lightning, New Text-to-Speech Model appeared first on Analytics India Magazine.
