HeadsUpAI

HeyGen Adds Custom Motion Cues to Direct Avatar V Performances

HeyGen, an AI video generation platform specializing in digital twins, launched Custom Motion to give users manual control over AI presenters. The feature works with Avatar V, which creates digital twins from short reference videos. Instead of relying on AI to predict movement, you can now type specific cues for gestures and expressions.
Feature name
Custom Motion
Supported model
Avatar V
Input method
Text-based cues
Example motions
Lean-in, smile, thumbs up, and more
Availability
HeyGen web platform (Presenter mode)

This update addresses the lack of intentionality in AI video. While previous models achieved visual realism, they often lacked the emotional timing required for storytelling. By enabling manual direction, HeyGen bridges the gap between talking heads and dynamic performers, similar to how Seedance 2.0 enabled avatars to interact within cinematic scenes.

You can now insert cues like a lean-in for key points or a thumbs up at the end. These controls are available within the Presenter mode on the HeyGen web platform. This brings professional-grade performance control to standard script-to-video workflows used for training, explainers, and marketing.

View the full update on app.heygen.com
HeyGen
HeyGen
@HeyGen
X

Avatar V now takes direction Custom Motion let you cue specific motions, expressions and gazes ā†’ A lean-in before a key point ā†’ A smile for emphasis ā†’ A thumbs up at the end Whatever the moment calls for, just type it https://t.co/lXaktehDgG

25retweets168likes
View on X

Still wondering? A few quick answers below.

Custom Motion is a new feature for HeyGen AI avatars that allows users to direct specific movements and expressions using text-based cues. Instead of relying on the AI to automatically generate gestures, creators can now type directions into their script to trigger intentional actions like nodding, smiling, or leaning in at specific moments.

Custom Motion is specifically designed to work with Avatar V, which is HeyGen's most advanced model for creating realistic digital twins from short reference videos. While HeyGen offers other modes like Cinematic for short clips, Custom Motion is integrated into the Presenter mode to improve the performance of longer-form talking head videos.

Custom Motion is accessed through the Presenter mode on the HeyGen web platform. While editing your script, you can insert text-based cue blocks at any point in the timeline. The AI avatar then performs the described action at that moment in the video, giving you the same level of directorial control a video director has over a human presenter.

HeyGen Custom Motion supports a wide range of expressive gestures and facial expressions. Examples shown by HeyGen include leaning forward before a key point, smiling during emphasis, and giving a thumbs up at the end. The system is designed to interpret natural language descriptions, meaning you can describe an intended expression and the avatar will attempt to perform it.

Yes. Standard HeyGen Avatar V automatically generates movement based on what it predicts will look natural for the spoken content. Custom Motion gives you manual override capability, letting you specify when and what gestures occur. This is distinct from the automated and random-feeling movement generated by most talking-head video tools, giving you predictable, reproducible results for professional productions.

Share this update