Real-Time AI
Conversations
Voice, text, and video input with generated speech and a rendered Unreal character, exposed through browser and server SDKs.
Quick Start
Connect with browser-safe credentials created by your backend
Loading...What You Get
Browser SDK
Connect to the session gateway, start voice or text calls, and embed the signed renderer player.
Server SDK
Create sessions on your backend and return browser-safe credentials to the client.
Usage Dashboard
Review sessions, usage, billing, API keys, and recorded performance telemetry.
Dedicated Runtime
Each active session is assigned its own conversational and rendering runtime.
Unreal Rendering
Runtime MetaHuman Lip Sync drives a hosted UE character streamed to the browser over WebRTC.
Session Security
API keys stay server-side while browsers receive time-limited session credentials and signed player URLs.
Cloud UE Characters
Run premium characters in the hosted renderer and stream WebRTC video back to the browser
Speech-synced facial motion
Cloud renderer container
Video and audio player stream
What the Platform Provides
Integrate
- Browser and server TypeScript SDKs
- Server-created session credentials
- Authenticated WebSocket session gateway
- Signed renderer player URL
- React hook and direct client APIs
Observe
- Active and historical sessions
- Usage and billing records
- Latency and error telemetry
- Scoped API-key management
- Session cleanup and expiry state
Run the Pipeline
- ASR and voice activity detection
- llama.cpp language-model serving
- Qwen3-TTS speech generation
- Runtime MetaHuman Lip Sync
- UE Pixel Streaming over WebRTC
One Browser-Safe Session Contract
Your backend creates the session. The browser receives only the gateway credentials and signed renderer URL it needs.
Simple Pricing
One plan, metered by session time
- No monthly fee or commitment
- Usage based on active session duration
- Up to 5 concurrent sessions
- Rate limits configurable to the plan ceiling