AI that talks like a human

A digital human that converses naturally over live video and voice.

Try the demo Contact us

1.5-2s

First response

260ms

Live transcription

25fps

Zero-dropout lip-sync

Languages supported

Production

Sona Persona, talking back in real time

A digital human that speaks like a person, over live video and voice.

Fast first response
Starts responding in ~1.5s for text and ~2s for voice.
Accurate lip-sync
Trained in-house for natural mouth shapes.
Native Korean voice
Natural speech tuned specifically for Korean.

Explore Sona Persona

Tech Preview

3D MetaHuman, real-time lip-sync

Unreal Engine MetaHuman-based 3D digital human.

Explore Sona MetaHuman

Ask, and a face answers

You speak

No install, no signup. It starts right in the browser with just a microphone.

It understands, then composes

It reads the intent of your question, then responds in ~1.5s for text and ~2s for voice.

A face replies

Self-trained lip-sync shapes natural mouth movement, streamed as live video.

From kiosks to museums

Six frontlines, one platform, a different face for each.

See the use cases

Customer kiosk

24/7 in-store & lobby guidance.

Education

1:1 tutoring & language learning.

Healthcare info

Reception & wayfinding in hospitals.

Museum & public info

Exhibition docent & tourism.

Retail

Styling recommendations & in-store guide.

Entertainment

AI character performance & interactive media.

Answers around the clock, data only for the session

Designed around public-sector and finance-grade security requirements.

On-premises option

Runs fully inside your network once GPU infrastructure is in place.

Discarded at session end

Conversation data is discarded the moment a session ends, by default.

Standard rollout in 1-2 weeks

From requirements to on-site deployment, on a standard setup.

Reply within 2-3 business days

Adoption inquiries get a direct reply within 2-3 business days.