APT MV AI Prompt

You do not need to write an apt ai video prompt

Most searches for an apt mv ai prompt come from people using a general video model, where the prompt is the entire product. This tool takes the opposite approach. The scene, the lighting, the camera move, the choreography and the audio are already described and locked inside the template, so there is nothing to type.

That is deliberate. A prompt is a fragile way to describe a scene with two people, a specific stage and a specific beat. Small wording changes move the camera, change the framing, or drop one of the performers. A fixed template removes that risk.

What a working prompt would have to cover

If you did write one, a prompt that reproduces this clip would need to specify all of the following, in order:

1. Two identity slots. One for the singer at the microphone and one for the drummer at the kit, with the order fixed so each upload lands in the right role.
2. Identity from the photos. Face, hair, skin tone, outfit and accessories come from the uploads, while background, framing and expression do not.
3. Motion from the template. The singing posture, the drumming motion and the closing dance come from the source clip, not from either photo.
4. Expression from the template. The performance energy has to come from the clip, otherwise both faces default to a neutral expression.
5. Scene and camera locks. The pink stage, the microphone, the drum kit, the wide two-shot framing and the continuous camera move all stay unchanged.
6. An exclusion list. No extra people, no captions, no logos and no watermarks.

That is a long specification, and every line is load-bearing. Getting one of them wrong is the difference between a duo clip and a single face swap on a stranger's body.

How the template replaces the prompt

Because this product ships one prepared scene, the specification above is written once and stored on the server. When you upload two photos and submit, the generator sends the locked scene, the two reference images and the template video to the model together. You never see the prompt, and you cannot accidentally break it.

What you still decide

You keep the decisions that actually change the result: which two photos to use, which output ratio fits the destination, and which resolution matches the final channel. Everything else is fixed.

A practical prompt alternative

If you are comparing tools, judge them on the output rather than on prompt length. The useful question is whether both performers are replaced, whether the audio survives, and whether the clip keeps the same beat as the source. Those three answers tell you more than any prompt template.

Try the prepared template instead of writing one.