Persona Studio

Create/edit a persona, manage photos + voice samples, and capture photo/audio directly in browser. This replaces the old separate Personas / Voice Profiles / Avatar Profiles operation.

Persona List

Create / Edit Persona + Voice + Avatar

No row selected.

3D Avatar Generator · MetaPerson

Generate a true 3D avatar from a photo, export GLB, then Hologram Agent will render the GLB with Three.js. REST API is Enterprise-only; this uses the official iframe + JS message flow.
MetaPerson keys are read from server-side configuration. They are no longer displayed on this page.
Open Creator after selecting a persona with a clear front photo. Creator is embedded inline; use buttons above.
MetaPerson StatusWaiting.
Use external buttons: Generate From Active Photo → Export GLB → Save Exported GLB. The iframe progress and generated model remain visible.

Assets: Photos + Voice Samples

Select a persona first.
Upload a voice sample, then prepare clone profile.

Captured / Uploaded Assets

No assets loaded.

Live Capture

Camera photo

Voice recording

Ready to record a voice sample.

Voice Clone Requirement

A persona can have multiple photos and voice samples. The latest uploaded or recorded voice becomes the active speaker reference. Click Prepare Voice Clone after uploading a voice sample. It calls the existing /api/voice.cfm flow, registers the latest voice sample with MiniMax through the local Python bridge, saves provider_voice_id, and generates preview audio. If provider fails, the UI will show fallback instead of pretending success.

cd C:\inetpub\demos\demo_hm\voice_clone_service
venv313\Scripts\activate
set HM_VOICE_PROVIDER=minimax
set MINIMAX_REGION=cn
set MINIMAX_API_HOST=https://api.minimax.chat
set MINIMAX_API_KEY=your_key
set HM_PUBLIC_BASE_URL=http://demos.e-xanke.com/demo_hm
python server.py --project-root C:\inetpub\demos\demo_hm