Jean Phil video prompt examples
Three starting prompts for dance, playful boxing and a café scene.
Room dance
An original fictional gentleman performs a small side-step dance in a quiet room. Keep his full body and feet in frame, a serious expression and a steady camera. One continuous shot. This is a suggested prompt, not a separately tested output.
Playful boxing
An original fictional character throws two slow shadowboxing jabs in an empty courtyard, then lowers his hands. No opponent or contact. Steady medium-wide shot. Choose a matching movement reference; prompts cannot guarantee exact timing.
Café scene
An original fictional gentleman on a café terrace makes a small bow, then looks toward the camera with a serious expression. Soft daylight, one continuous medium shot. Change one instruction at a time; every new generation consumes the quoted credits.
Know which mode receives your words
Write a prompt sends the text you enter and the selected movement reference, without a photo. Character image uses the preset plus your optional additional direction and the uploaded image. A sentence such as “keep my hairstyle” can conflict with a preset that asks for a blond bob. More text does not reliably resolve conflicting instructions.
A proposed identity-first variant — not tested
Proposed instruction: “Use the uploaded image as the only identity and appearance reference. Keep its face, hair, facial hair, skin tone and clothing. Use the attached video only for body movement and timing; do not copy its performer’s face, hairstyle or outfit. One continuous full-body shot.” This is a candidate for a future photo-mode preset test, not a validated recipe or a hidden switch in the current interface.
How to make a meaningful comparison
Use the same authorized image, reference video, duration, resolution and format for both versions. Change only the appearance instruction, record the full prompt and credit cost, then compare identity, clothing, motion, artifacts and waiting time. There is no published A/B result for the candidate above. The two existing room-dance videos are generation evidence, not a prompt benchmark.
What has been tested, and what has not
| Version | Appearance instruction | Evidence |
|---|---|---|
| Current photo preset | Preserve facial identity, but add a blond bob, moustache and suit. | Two room-dance outputs; appearance changed. No controlled A/B test. |
| Identity-first candidate | Image controls identity and clothing; video controls motion only. | Not generated or quality-validated. No result comparison is available. |