Skip to content

TANGO Model

Co-speech gesture video reenactment from audio

Visit official site

What TANGO Model is

The technique is described as co-speech gesture video reenactment built on hierarchical audio-motion embedding and diffusion interpolation. It maps speech signals to hand gestures in a video.

What you can do with it

  • Give film characters natural hand gestures
  • Generate gestures for game NPC characters
  • Add gestures to an instructional video

Getting started

  • Visit the site and get the code and API
  • Read the docs and examples for integration
  • Prepare a reference video and a target speech input

What to watch out for

  • The speech signal needs to match the gestures
  • Integration is handled by the developer

Pros & cons

✓ What we like

  • Code and an API are provided for integration
  • Multiple devices and operating systems are named

! What to watch out for

  • A reference video and a clean speech input are needed
  • Integration is done by the developer

FAQ

Who developed it?

The University of Tokyo and CyberAgent AI Lab are named.

What is required as input?

A reference video and a target speech signal.

Last reviewed: 2026-11-15

More AI video generation tools

View all →

How we review