Archive
A Deterministic, Deficit-Aware Framework for Egocentric Data Collection
Published September 1, 2026
Read article →Full duplexListen + speak
Animated spectrogram for full duplex voice.01 / Company
Bellu Technologies Private Limited · Madanapalle, Andhra Pradesh, India · Est. 2025
Full duplex AI.
Bellu.ai is an operating company. We are building full duplex voice: expressive, emotional, with non-verbal support and 22 Indian languages.
/ Full duplex
Listen and speak on the same stream. No wait, then talk.
/ Barge-in
Interrupt mid-sentence. Yield is native, not a VAD hack.
/ Expressive
Emotion and tone live in the voice, not a filter after speech.
/ Non-verbal
Breath, pause, overlap, and backchannel stay in the model.
/ 22 Indian languages
Scheduled Indian languages as first-class, not a translation layer.
/ One clock
Hearing and speech share one timeline. Latency does not stack.
02 / Team
Founder
Named, reachable, and linked to public third-party profiles on this page.
Founder · Madanapalle, Andhra Pradesh, India
Santhosh
Founder of Bellu Technologies Private Limited.
santhosh@belluai.com03 / Research
Lab notes
Open research folder →Archive
The Architecture of Embodiment: Multimodal Data Standardization and Management for Generalizable Embodied AI
Published November 14, 2025
https://www.techrxiv.org/users/998761/articles/1360379-the-architecture-of-embodiment-multimodal-data-standardization-and-management-for-generalizable-embodied-ai04 / Contact
Bellu Technologies Private Limited. Direct to the founder.
santhosh@belluai.comMadanapalle, Andhra Pradesh 517325, India