Spatius
Animates lip-synced 3D digital-human faces from audio in real time, affordable even on low-power devices
Last verified:
What is Spatius?
Spatius is a digital human AI platform that generates real-time, lip-synced 3D facial animations from audio streams using a cloud-edge hybrid rendering architecture. It enables affordable AI avatars on constrained networks and low-power devices, designed for language learning, HR interviews, mental health companions, in-vehicle systems, and AI hardware.
Spatius pricing
Pricing model: Freemium
Free tier available; AI avatar time at $0.42/hour. Spatius Scale Plan at $0.007/min (vs. ~$0.15/min industry average). Quick start in 10 minutes.
Spatius pros
- Extremely affordable pricing: $0.42/hour for AI avatar time; $0.007/min with Scale Plan (20x cheaper than ~$0.15/min industry average)
- Runs 1080p at 25fps on entry-level chipsets (G88, S565, 8189, RK3576) without dedicated GPUs; deploy on phones, monitors, laptops
- Direct audio-to-avatar engine with real-time lip-syncing; supports LiveKit WebRTC and WebSocket protocols
- Custom 3DGS avatar integration for proprietary avatars; built-in payload redundancy for network stability
Spatius cons
- Custom avatar integration marked as 'Beta Access' — may have limited stability or incomplete features
- No performance metrics, latency benchmarks, or animation quality details provided
- Limited documentation on language support, API complexity, and integration requirements beyond LiveKit
Frequently asked questions about Spatius
How much does Spatius cost compared to alternatives?
Spatius Scale Plan costs $0.007/min versus ~$0.15/min for traditional cloud streaming — approximately 20x cheaper for the same AI avatar time.
What devices and chipsets can run Spatius?
Runs 1080p at 25fps on entry-level chipsets (G88, S565, 8189, RK3576) without dedicated GPUs. Deployable on phones, monitors, laptops, in-vehicle systems, and kiosks.
How does the audio-to-avatar engine work?
Accepts audio streams via LiveKit WebRTC or WebSockets and outputs real-time, lip-synced 3D facial animations using cloud-edge hybrid rendering with a lightweight 100 kbps stream.
Can I integrate custom avatars?
Yes, Spatius supports custom-trained 3DGS models for proprietary avatars, available in beta access.