Pixa AI, a Jaipur-based artificial intelligence startup, has launched Luna, which marks India’s first speech-to-speech foundational voice model that enables real-time, emotion-driven interactions without text intermediaries.

Founded by 25-year-old IIT-BHU alumnus Sparsh Agrawal in 2024, the company has raised an undisclosed seed round from Cred founder Kunal Shah, former Carlyle Asia Growth Partners MD Shankar Narayana, actor-investor Kunal Kapoor, and Nikhil Kamath’s WTFund.

Luna claims to differ from existing Indian voice AI solutions like Gnani and Sarvam by operating as a pure speech-to-speech system rather than routing through text conversion layers. This architecture, the company claims, enables sub-600-millisecond end-to-end latency.

“The future of voice AI isn’t automating call

See Full Page