About Soltio

LIVA is a real-time avatar engine that makes AI conversations feel face-to-face. Avatars render on the client (any device or browser), so you don’t need cloud GPUs.

Chatbots type. LIVA talks, gestures, and reacts. You get lifelike lip-sync, expressions, and non-verbal cues—so the conversation feels human, not scripted.

A patent-pending interactive-streaming protocol renders video locally on users’ devices. That cuts infrastructure costs dramatically and keeps latency ultra-low.

Bring your own: OpenAI, Google Gemini, Anthropic, Mistral, Cohere, local endpoints—mix and switch as you like. LIVA is model-agnostic.

Yes. LIVA runs in modern web browsers and inside mobile apps via WebViews today. A React Native SDK is planned, with native iOS/Android SDKs to follow on the roadmap.

Sub-second end-to-end in typical conditions (network dependent). The goal is to keep it “blink-fast” so turn-taking feels natural.

No. Rendering happens client-side. Your only server costs are your LLM/TTS usage and routine API calls.

Yes. Add PDFs, docs, FAQs, or URLs. LIVA indexes your content so avatars answer with brand-accurate information. Update it anytime.

Completely. Theme outfits, backgrounds, camera framing, gestures, tone, and safety level. Save presets per product or market.

Avatar & voice cloning is coming soon. You’ll be able to turn a single photo and short audio sample into your digital twin. (Early access waitlist available.)

LIVA supports multiple major languages and a range of TTS providers (e.g., ElevenLabs). Pick voices per locale; swap providers as needed.

Rendering is local to the device. You choose where model calls go (your preferred LLM provider). Conversation storage is opt-in. Enterprise controls (logging, redaction, data retention) are available.

Avatars render locally, but most real-time experiences need a network connection for LLM/TTS. You can cache scripts for limited offline flows.

Because rendering happens on users’ devices, LIVA scales horizontally with your audience. Serve thousands of simultaneous conversations without GPU farms.

Add a few lines of JS to embed an avatar widget, point it at your model endpoint, and (optionally) connect your knowledge base. Sample projects and a quick-start guide are provided.

For developers, pricing is usage-based (per-token/voice minutes). Businesses can license custom avatars and enterprise features. The consumer app offers a free tier with an optional premium subscription. (See the pricing page for current details.)

Yes, with the right guardrails. You can enforce content policies, routing, redaction, and human-handoff. For regulated deployments, talk to us about compliance needs.

Yes. For example, LIVA powers a smart concierge in Dubai taxis, helping riders get instant guidance during trips.