Hello,
We are building Care Navi Health, a real-time health navigation assistant for patients (web + mobile). It runs on Gemini 3.8 Live (Live Avatar) over Vertex AI, and works in production today with a prebuilt avatar.
We would like to request allowlist access to the CUSTOM AVATAR feature (avatar_config.customized_avatar) so we can use our own branded caregiver avatar instead of a prebuilt one.
Current behavior when we send a customized_avatar in the setup payload:
Google closes the session with 1007
“Current project is not allowlisted for customized avatar feature.”
Details:
- Project ID: parkpin-507914
- Billing: enabled (active billing account 011DE2-151D91-BE1…)
- Location: us-central1
- Model: projects/parkpin-507914/locations/us-central1/publishers/google/models/gemini-3.8-live
- Endpoint: wss://us-central1-aiplatform.googleapis.com/ws/google.cloud.aiplatform.v1.LlmBidiService/BidiGenerateContent
- Feature: Live Avatar — customized avatar (reference image, 9:16 PNG)
- Paired voice: prebuilt HD voice (Leda)
Use case: a patient-facing health navigation assistant (hospital/ER/pharmacy directions, medical questions). We use a single branded caregiver avatar image (a nurse illustration) that is our own production asset; no end-user face samples are processed, and the reference image is under our control and consented.
We have a working integration and can provide additional technical details on request. Thank you!