Dr. Wilhelm | our technical explainer
Sparkle Cloud focuses on AI companion technology, building interaction experiences that move beyond mechanical responses. By combining LLMs with emotion recognition, our AI engages in deep, emotionally aware conversation responsive to the current dialogue context.
Through TTS (Text-to-Speech) and ASR (Automatic Speech Recognition), we deliver natural, fluent voice interaction with personalized AI voices. Long-term memory and context-adaptive learning let the AI tune its style to each user, producing proactive responses closer to real interpersonal relationships.
Whether for elderly companionship, education, livestream promotion, or social practice — our technology ensures AI characters are smarter, more natural, and emotionally engaged.
Four layers carry every warm conversation, from multi-modal interaction down to cloud infrastructure:
From the MOEA "Proactive AI Smart Health Management Service" program, validated across day-care centers, community care stations, and home care:
90%+
Multi-modal dialogue success rate
95%+
Health-data reminder success rate
100%
Field service satisfaction
10%+
Caregiver efficiency gain
Sparkle Onyx is Sparkle Cloud's AI companion interaction system, designed for VTubers, AI companion education, and virtual livestreaming. Multi-modal AI — image recognition, ASR, LLMs, Live2D expression driving, and emotion recognition — gives AI characters fluid, natural, personality-rich interactivity.
Sparkle Onyx provides advanced AI memory: it adapts conversation style to each user's history and interests, supports real-time voice dialogue and personalized voice synthesis, and brings AI characters closer to real human interaction. Virtual idols, interactive education, elderly companionship, brand ambassadors — Sparkle Onyx delivers a best-in-class AI interaction experience.