Omli raises $1.5M seed to build AI that understands how kids speak
What's the deal? Omli, a Gurugram-based startup, has raised $1.5M in seed funding to develop AI tools designed specifically for how children speak. Its app, OmliKids, features an AI companion named Doro that helps children aged three to 10 practise speaking and thinking skills through voice-based chats.
The app has already crossed 500,000 downloads on Google Play within a few months of launch.
Why now? Most speech recognition technology is built for adults and struggles with children's voices — especially those speaking with regional Indian accents or in local languages. Omli built its own model, VedAnsh-v1, to close that gap.
"The ambition is to bring children's speech recognition close to adult-grade accuracy, including Indian accents and vernacular languages," said co-founder Siddhant Garg.
What could go wrong? Building AI for young children raises the stakes on safety, privacy, and accuracy. If the speech model misinterprets a child's words or fails to engage them, parents will move on quickly. Competing with well-funded edtech players — and keeping kids' attention — is no small feat either.
The signal: AI tools are moving beyond general-purpose adult applications and into specialised niches like children's education. Omli's early traction suggests real demand for voice-first learning tools tailored to young users in multilingual markets like India. The startup also plans to develop features that detect speech issues early, potentially expanding into health-adjacent territory over the next six months.
Read more: NewsBytes
