Spatius

Spatius: Real-Time AI Avatar Infrastructure

Spatius provides real-time AI avatar infrastructure designed to create next-generation digital humans. By utilizing a cloud-edge hybrid architecture, the platform replaces heavy cloud rendering with a lightweight 100 kbps stream, rendering directly on the edge to significantly cut infrastructure overhead.

Key Features

• Direct Audio-to-Avatar Engine: Accepts audio streams via standard protocols like LiveKit WebRTC or WebSockets to output real-time, lip-synced 3D facial animations without requiring external animation pipelines.
• Deploy Anywhere: Features a deployment architecture ready for phones, monitors, and laptops.
• 1080p On Entry-Level Chipsets: Runs 1080p at 25fps directly on entry-level chipsets (such as G88, S565, 8189, RK3576) without the need for dedicated GPUs.
• Low Compute Overhead: Includes built-in payload redundancy to maintain stable playback during network jitter.
• Custom Avatar Integration (Beta): Supports the deployment of custom-trained 3DGS models, allowing users to turn a single photo into a custom AI avatar and integrate proprietary avatars alongside free stock avatars.

Use Cases

• Language Learning: Conversational language tutors with stable voice-driven avatars on constrained networks.
• Interviewers & HR Tech: Digital avatars for human resources and interviewing applications.
• Companions & Mental Health: Avatar interfaces for companionship and mental health support.
• In-Vehicle & Kiosks: Integration into automotive systems and interactive kiosks.
• AI Hardware: Deployment on dedicated artificial intelligence hardware devices.

Pricing

Spatius offers a Scale Plan utilizing a cloud-edge hybrid model priced at $0.007 per minute, which equates to $0.42 per hour.

Getting Started

Website: https://www.spatius.ai/

Spatius enables the deployment of affordable and accessible digital humans across various industries through efficient edge rendering technology.