Avatar Studio

What is Avatar Studio

Avatar Studio lets you create personalized talking head avatars. You record or upload your voice (~30 seconds), provide a front-facing photo, and the system generates a talking head video — your avatar speaking synchronized with your cloned voice.

Once created, avatars are saved in your library and you reuse them as many times as you want: just provide a script, and the avatar will generate a new video in the background.

How to use it

Avatar Studio is organized into three tabs:

1. "Create" tab

Here you record a voice and upload a photo to create a new avatar.

Step 1: Record or upload your voice

• Click "● Record" and read aloud the provided text (about 30 seconds). The live equalizer shows your microphone capturing sound.
• Or click "⬆ Upload audio" and select an audio file of your voice.
• Give the voice a name (e.g., "My voice") and click "Clone voice". Cloning happens in the background.
• Voice preview: you can listen to your cloned voice with any text using the "▶ Play" button before proceeding.

Step 2: Snap or upload a photo

• Click "📷 Webcam" to snap a photo with your webcam, or "⬆ Upload" to select a photo from disk.
• The photo should be front-facing and well-lit: it will become your avatar's face.

Step 3: Save the avatar

• Give the avatar a name (e.g., "Myself").
• Choose the engine:
• SadTalker: sharp lip-sync, single photo.
• MuseTalk: natural lip-sync, accepts photos or videos.
• OmniAvatar (local): full-body motion on your GPU, free but slow (~3 min per video second).
• OmniHuman (premium): full body and head animation, looks like a real human.
• VEED Fabric (premium): photo transformed into realistic video.
• Select the cloned voice you created (optional: you can skip and use an existing voice).
• Click "💾 Create avatar".

2. "Library" tab

View all saved avatars and voices. Here you can:

• See each avatar with thumbnail, name, and associated voice (✓ if present, "no voice" otherwise).
• Delete avatars or voices you no longer need.

3. "Generate" tab

Create videos with your saved avatars.

• Select an avatar from the dropdown.
• Write the script — the text your avatar will speak.
• Enable "Karaoke caption" to display the text as synchronized subtitles.
• Click "🎬 Generate avatar video".

Generation happens in the background queue. You'll receive a desktop notification when the video is ready. The video appears in the Gallery and you can download it from there.

Options and premium features

If a premium engine button shows a lock, premium quality is not enabled on your account.

Tips

Voice quality: Record in a quiet environment without background noise. Clear, natural speech results in better cloning.

Optimal photo: Use a well-lit, front-facing photo with your full face visible. Avoid sunglasses or hats that cover your face.

Reuse avatars: Once saved, avatars stay in your library. You can generate dozens of different videos with the same avatar and different voices, without recreating anything.

OmniAvatar local: Free but slow. Use it for background rendering when you can wait; for fast results, choose a premium engine.

Common issues

"Voice cloning failed"

• Your recorded voice may be too short or poor quality. Record at least 30 seconds in a quiet environment and try again.

"Avatar not found"

• Make sure you've saved at least one avatar in the "Create" tab before generating.

"Buttons don't work" / nothing appears

• Clear your browser cache: press Ctrl+Shift+R (Windows/Linux) or Cmd+Shift+R (Mac).

Desktop notifications don't arrive

• Make sure your browser has permission to show notifications. In some browsers, notifications require the site to be added to your home screen or opened in a dedicated tab.