A photo becomes a speaking presenter
HeyGen supplies software for creating avatar videos from images, scripts and speech. The product combines a voice track with generated facial movement and visible delivery, allowing a team to create presenter-style material without recording a new camera performance for each version. [1]
Its current Avatar IV interface supports lifelike and stylized characters, with expression and gesture controls. The company separates the original photo-based model from its later Digital Twin feature, which its product chronology places in August 2025. Those are related capabilities with different release histories. [1]
The Avatar IV rollout
HeyGen CEO Joshua Xu’s May 6, 2025 Avatar IV announcement described an audio-to-expression engine that used a voice's rhythm and tone to generate facial dynamics. It announced a rollout to all users, starting from one photo with a script or voice input. [2][3][4]
The important contribution is a controllable connection between speech and appearance. Describing a model as responding to emotion concerns the generated performance; it does not show that the avatar experiences that emotion. Later product pages should not be used to assign every current control to the original May release. [2][1]
Sources
- Create talking photo avatars in 1280p+ HD resolution
HeyGen · retrieved Oct 23, 2026
- HeyGen Avatar IV launch thread by CEO Joshua Xu
Joshua Xu · May 6, 2025
- HeyGen Launches Avatar-IV: AI Avatars With Realistic Gestures and Voice Sync
AiNews · May 7, 2025
- Joshua Xu
HeyGen · retrieved Oct 23, 2026