1. Choose the performance and frame
Use restrained shadowboxing for the most recognizable visual cue, a runway walk for editorial motion, or a small disco sway for a dance interpretation. Choose 9:16 for a vertical social post, 16:9 for a landscape shot, or 1:1 for a square portrait.
Start with a single character and a static camera. Add a setting that supports the action without crowded backgrounds or text.
2. Build and review the portrait
In the prompt studio, choose a performance and write your setting. “Build my prompts” prepares both portrait and motion text without contacting an AI provider.
Once live rendering is configured, sign in and generate the portrait. Check the face, bob, moustache, suit pattern, fingers and feet. Approve the portrait only when it is a useful starting image.
3. Animate the approved image
The video stage uses the approved image task, not an arbitrary image URL. Our configured implementation uses Nano Banana Pro for the still image and Kling 2.6 for a five-second silent clip. These are our implementation choices, not verified models used by the original creator.
An asynchronous provider task is checked for a real result. A waiting or failed task is not presented as a completed video. Keep your session so you can resume the last task from the studio.
4. Edit, label and share
Watch the full clip for face drift, hand distortions and unexpected movement. Use a video editor to trim, add captions and attach audio you have permission to use. The studio does not bundle a viral soundtrack or promise a seamless loop.
Disclose AI-generated or AI-assisted media as appropriate for the platform. The photo mode guides the portrait appearance; motion-video mode uses Kling motion control and may retain reference audio. Only upload files you own or have permission to use.