1. Start with a usable portrait
Use a sharp image with one person, visible face, and even light. A camera-facing head-and-shoulders photo gives the model more stable identity cues than a group shot, a profile, or a heavily filtered selfie.
- Use a JPEG, PNG, or WebP under 10 MB.
- Avoid hats, hands, microphones, and sunglasses covering the face.
- Only upload yourself or a person who has explicitly agreed.
2. Pick the Bulin 47 template
The template supplies the setting, camera energy, vertical framing, and performance rhythm. Your upload is used as the person in the clip. That is why a template is different from a generic text-to-video prompt: it aims for a repeatable trend format.
3. Review before sharing
AI video can change small facial details between frames. Watch the rendered clip, download only the version you are comfortable publishing, and label it as AI-made when the platform or audience context calls for it.
What if my result does not look right?
Try a new photo before changing everything else. Better source images usually produce a bigger improvement than longer instructions. Keep the person centered and leave a little room above the head.