Realtime avatar audio retention, training, and custom SLA

General
1

Hi Protoface team,

We’re integrating Protoface Realtime into a B2B voice-training product via the LiveKit plugin, and so far the quality is great. Our voice agent streams generated speech audio to Protoface to drive lip sync; we are not uploading audio files as assets. A few questions before we roll it out to customers:

Could you clarify:

  1. Is that realtime audio recorded or stored anywhere, including logs, debugging systems, or backups? If so, for how long, and how can we request deletion?

  2. Are the audio stream or any data derived from it used to train or improve models by Protoface or any third party? Can we obtain a written no-training commitment?

A link to the relevant policy or agreement would be helpful. Thanks!

2

Hi, thanks for the outreach!

On our public plans, for LiveKit integrated avatars, we can confirm that the audio is not retained and not used for training.

Please note that our policies for public plans are subject to change unless you are on an enterprise contract.

1 Like