CreationsFavoritesAssetsProjects
All
Image
Video
Features
Back to Agents

AI Talking Avatar

Upload a presenter video and enter a script to create a natural talking video with Seedance 2.5. Optionally add a product photo for the same presenter to showcase and introduce your product.

Required

Upload a single-person MP4 with a clear, unobstructed face: 2–30 seconds, up to 100 MiB. The model uses the person's appearance as a reference for a new talking video.

Optional

Optional: upload one clear JPG, PNG or WebP product photo, up to 10 MiB. The presenter will showcase the product while reading your script. Leave empty for a regular talking video.

Required
0 / 200
View my creations
Your next creation starts here

Complete the settings to see your generated results.

Frequently asked questions

What do I need to upload?

Provide a 2–30 second, single-person MP4 with a clearly visible face and the script you want spoken. A product photo is optional: one JPG, PNG or WebP, up to 10 MiB. Adding it asks the presenter to showcase and introduce the product; leaving it empty creates a regular talking video. Choose a 15, 20 or 30 second output. About 20–35 English words suits 15 seconds.

Will it preserve the original voice and video exactly?

Seedance 2.5 generates a new video and voice using your presenter as a reference. The prompt requests consistent appearance, exact script delivery and synchronized lips. Results can vary; the original voice, every frame and perfect lip sync are not guaranteed.

Drag the cat to move it. Click to contact support.