Hedra Omnia

Hedra Omnia is a character-driven video model that jointly processes a reference image, audio, and a written scene direction. The image establishes the subject, the audio supplies words, timing, rhythm, and emotion, and the prompt directs body language, expression, camera movement, lighting, and background behaviour. This unified approach lets a person perform inside a changing scene rather than remaining a face pasted onto a fixed frame. Omnia is available inside Hedra Studio with square, landscape, and vertical outputs at several resolutions.

The workflow reads like a compact directing brief: choose who is on screen, provide what they say, then describe how the moment should feel and how the camera should witness it. Quiet glances, pauses, gestures, and camera drift can make the result feel more authored than a traditional talking avatar. Current clips are short, generation takes time, and busy references, hands, logos, fast speech, or elaborate interactions can expose instability. Omnia is therefore best treated as a shot generator for ads, podcast moments, cinematic avatars, and social scenes rather than a system for carrying a long uninterrupted performance.