LiveKit V2V Translation Agent

Configure your real-time voice translation settings

Ready to configure
ℹ️ Language Settings Moved

Language preferences (source and target) are now configured per-participant on the client interface. Each user can choose their own source and target languages when joining a room.

Voice Settings

Choose the voice for translated speech output

System Prompt Configuration

Override the default translation system prompt. Leave empty to use auto-generated prompt based on participant language preferences. The prompt will be used by the translator with each participant's specific source and target languages.

System Prompt Template:

Loading...
Note: The actual prompt will use each participant's language preferences. This shows the template structure.

Input Audio Transcription

Model used for audio transcription

Turn Detection (Voice Activity Detection)

Voice activity detection method
0.5 Sensitivity of voice detection (0.0 - 1.0)
Audio buffer before speech detection
Duration of silence to detect end of speech

Audio Publishing Settings

Enable to publish the original audio (passthrough/echo) alongside the translated audio. Disable to only publish translated audio.

Model Parameters

0.8 Controls randomness in responses (0.0 = deterministic, 2.0 = very random)
Maximum length of generated responses

Configuration Preview (JSON)

{}