Maya Research, an AI startup building voice interface technology, has raised $1.9 million in a seed funding round led by South Park Commons. Its open-source voice model, called Maya 1, was released under the Apache 2.0 licence and currently holds a Quality Elo score of 1,051 on Speech Arena, placing it sixth among open-weight voice models globally and putting it roughly on par with OpenAI's GPT-Realtime.

Maya Research claims that it is currently the only Indian company listed on Speech Arena. Maya Research’s models have crossed 440,000 downloads on Hugging Face, and the company's consumer application has been downloaded more than 3 million times across India, Southeast Asia, and the Middle East and North Africa region. The consumer app creates a data flywheel for high-quality, hard-to-collect data, providing Maya Research with an advantage in creating models faster and better.

Dheemanth Reddy, Co-founder and CEO of Maya Research, said, "India is not a text economy, it never was. While the conversation around AI sovereignty has focused on large language models and text interfaces, the more urgent question is: whose voice models are Indians talking to? Right now, that answer is almost entirely foreign. Maya exists to change that. We built our models frugally, on cloud credits, because we believed the problem was worth solving regardless of resources."

About Maya Research

Maya Research has been founded in 2025 by BS Dheemanth Reddy and Bharath Kumar Kakumani. Maya is building a voice interface for people who interact with technology through conversation rather than typing, tapping, and navigating complex interfaces.