Wan Speech

What it can do

• Modes: Talking avatar (lip-sync).
• With sound: yes.
• Takes your voice track and drives the lips to it.
• Has a "what must not appear" field (negative prompt).
• Has a seed, so a good result can be repeated.

Modes and prices

The ⚡ price is calculated from the Studio price list on October 16, 2026: 1 ⚡ ≈ $0.025. The exact amount is always shown on the Generate button before you start; it depends on duration, quality and sound. Not every combination of duration and quality is available for every model: the upper price is an estimate at the maximum values.

What it is good for

• A talking face: photo + voice track gives a clip with synced lips.
• A clip with sound and spoken lines in one run.

How to pick it in the Studio

1. Open the Studio

   Go to gibrid.ai/studio and log in.

2. Choose Video

   Avatar models are grouped in the model list under Avatars.

3. Add a photo and a voice

   Put the photo into Start frame, then add a Voice track: your own recording or a ready voiceover from My recordings.

4. Pick "Wan Speech"

   Click the model name above the input field and find it under Avatars.

5. Check the price and start

   The price in ⚡ is shown on the Generate button before you start.

Prompt tips

• Avatar models take a photo and a voice track: the photo sets the face, the track drives the lips.
• You can take a ready track from My recordings (the Voiceover section) or upload your own: a ready voiceover drives the lips and motion.

Common mistakes

• Upload only photos of people who agree to such use.
• Do not use a track longer than the model limit: the limit is in the duration facts above.

Related models and tiers

FAQ