Skip to main content
The Spatius demo repository is the runnable companion to these docs. Use it when you want to clone code, fill credentials, and see an avatar render, connect, and speak.

spatius-ai/spatius-avatar-demo

One consolidated repository containing every demo referenced from these docs.
SDK Reference pages document the packages and APIs you import into your app. Demo projects are end-to-end examples that combine SDKs, credentials, client apps, and optional backend services.

Demo matrix

PlatformDirect ModeBackend ModeLiveKit Agents IntegrationAgora Convo AI Integration
WebWeb quickstart — bundled sample audio plus optional realtime conversation
Web reference clients — React / Vue / vanilla / Next.js
Web Backend Mode clientLiveKit Agents Integration quickstart — Gemini Live quickstart
LiveKit Agents Integration reference — cascade & end-to-end backends
Convo AI Agent guide, TEN extension source, and Agora client guide. Packaged Web demo not provided in spatius-avatar-demo today.
iOSiOS Direct ModeiOS Backend Mode clientNot provided today. LiveKit iOS SDKs exist, but the current AvatarKit RTC demo path is Web-only.AvatarKitRTC iOS SDK plus Agora client guide. Packaged iOS demo not provided in spatius-avatar-demo today.
AndroidAndroid Direct ModeAndroid Backend Mode clientNot provided today. LiveKit Android SDKs exist, but the current AvatarKit RTC demo path is Web-only.Android RTC SDK support via ai.spatius:avatarkit-rtc plus Agora client guide. Packaged Android demo not provided in spatius-avatar-demo today.
FlutterFlutter Direct ModeFlutter Backend Mode clientNot provided today. The current AvatarKit RTC demo path is Web-only.Not provided today. Agora Convo AI client support currently covers Web, iOS, and Android.
Classification reminder:
  • LiveKit Agents Integration belongs to Platform Integrations.
  • Agora Convo AI Integration belongs to Platform Integrations.
  • Direct Mode and Backend Mode belong to Standalone Integrations.
  • The LiveKit Room example validates the RTC Adapter wiring against a LiveKit room (token, connect, adapter init, avatar load, mic publish). Remote audio playback and motion-driven rendering only happen when a producer is publishing into the room — it is not the full Backend Mode + RTC transport voice-agent demo.
Transport options such as Backend Mode with LiveKit, Backend Mode with Agora egress, and your own WebSocket are configured inside the relevant integration guide. They are not top-level demo categories.
GoalStart hereWhat it proves
Web Direct Mode quickstartquickstartApp ID, Avatar ID, Session Token, bundled sample audio, optional browser realtime conversation, Motion Server connection, and avatar rendering.
Full Web client referenceWeb Direct Mode referenceDirect Mode SDK integration across React, Vue, vanilla JavaScript, and Next.js.
iOS / Android / Flutter Direct Mode samplesiOS Direct Mode, Android Direct Mode, or Flutter Direct ModeNative and Flutter AvatarKit setup with runtime Session Token input.
Fastest Web voice-agent demoLiveKit Agents Integration quickstartLiveKit token server, agent dispatch, livekit-plugins-spatius, AvatarKit RTC, and microphone flow.
Full LiveKit Agents Integration referenceLiveKit Agents Integration referenceCascade and end-to-end agent backends with Vite and Next.js frontend apps.
Agora Convo AI IntegrationConvo AI Agent guideConfigure Spatius as the Convo AI avatar provider, then pair it with the Agora client guide. Use spatius_avatar_python only for direct TEN Framework graphs.
Backend-owned audio pipelineBackend Mode referenceBackend-owned ASR, LLM, TTS pipeline and thin clients that consume encoded audio and motion messages.
RTC Adapter LiveKit exampleplatform-integrations/livekit-room-demoValidates @spatius/avatarkit-rtc + LiveKitProvider wiring (token, connect, adapter init, avatar load, mic publish). Remote audio playback and motion rendering only when a producer is in the room. Not the full voice-agent pattern.
Backend Mode + RTC transport (full voice agent)Not packaged in this repository yetFull pattern: backend owns ASR / LLM / TTS / Server SDK, mints LiveKit tokens, and publishes audio + motion into the LiveKit room; RTC Adapter renders in the Web client. Assemble from the Backend Mode demo + the LiveKit example above, or use the LiveKit Agents quickstart for an end-to-end runnable equivalent.

Clone and run

git clone https://github.com/spatius-ai/spatius-avatar-demo.git
cd spatius-avatar-demo
Each demo folder has its own README with .env.example, dependency installation, and run commands. The READMEs are the source of truth for run instructions; this page is the index.

Next steps

Choose your integration path

Pick a Platform Integration, Direct Mode, or Backend Mode.

Quickstarts

Run the platform SDK quickstarts before diving into a full demo.