LipSync v4 Stream (Buffered)

Stream mode WebSocket URL JWT Token (optional) Avatar ID Voice clone ID Provider name Request ID (session) Voice language Player Layout

Audio Engine: Suspended

Text to speak (TTA)

Send text over WebSocket; server streams voice via WebRTC chunk-by-chunk.

Paragraph:

Words/chunk: Interval (ms):

Calls QwenVLLMServer.stream_chat over Modal RPC, streams the answer through TTS + LipSync over WebRTC.

Question:

Project ID:

Client ID:

LLM Module:

Source Lang:

Target Lang:

Avatar Identity (optional — activates persona):

Gender (optional):

Session ID (optional — for multi-turn history):

Audio Source (file)

Audio File

Ready

Status: Disconnected