properties.avatar block.
If you run a TEN Framework graph directly instead of starting a managed Convo AI agent, use the TEN Extension guide.
Token usage
Agora RTC tokens are used in three places when Spatius avatar output is enabled:| Token | Where it is used |
|---|---|
| Agent RTC token | Used by the Convo AI agent so it can join the Agora channel. |
| Avatar RTC token | properties.avatar.params.agora_token so Motion Server can publish avatar output into the same channel. |
| Client RTC token | Passed to the Web, iOS, or Android client so the user can join the same channel. |
Spatius avatar provider
Addproperties.avatar to the Convo AI agent configuration:
Avatar fields
| Field | Required | Description |
|---|---|---|
properties.avatar.enable | Yes | Set to true to enable avatar output for the Convo AI agent. |
properties.avatar.vendor | Yes | Set to spatius. |
properties.avatar.params.spatius_api_key | Yes | Spatius API Key. Keep it server-side. |
properties.avatar.params.spatius_app_id | Yes | Spatius App ID. |
properties.avatar.params.spatius_avatar_id | Yes | Avatar ID to render. |
properties.avatar.params.agora_uid | Yes | Agora UID used by the avatar publisher. Must be distinct from the user and agent UIDs. |
properties.avatar.params.agora_token | Yes | RTC token for agora_uid in properties.channel. |
properties.avatar.params.region | No | Spatius region. See Regions. |
properties.avatar.params.sample_rate | No | Avatar speech audio sample rate. Use 24000 unless your Convo AI setup requires a different value. |
properties.avatar.params.session_expire_minutes | No | Spatius server session lifetime. |
Next steps
Client
Join the Agora channel and render avatar output with
AgoraProvider.TEN Extension
Use
spatius_avatar_python when you run a TEN Framework graph directly.