Charme Dance AI Prompt: Describe the Groove, Camera and Scene

A useful Charme dance AI prompt describes a relaxed dance, a readable subject and a simple camera shot. It should not ask the model to reproduce somebody else's viral video frame for frame. On Charme Dance AI, the built-in direction establishes the dance scene. Your optional instructions help specify the mood, clothing or setting. Text mode creates a fictional adult; image mode uses one consenting adult photo as a visual reference.

The default request is a twelve-second, vertical 720p video using standard Seedance 2.0. It costs thirty credits, displayed before submission. Preparing or editing your prompt does not submit a paid generation. The tool embedded below lets you prepare either input mode on this page.

Start with the movement, not a list of camera effects

Charme belongs to a Brazilian social dance tradition associated with soul and R&B. For a short generated scene, describe small side steps, relaxed knees, a gentle shoulder groove and hands held naturally near the waist. These instructions are easier to reconcile than simultaneous demands for acrobatics, rapid spins and perfect facial stability.

A prompt is not a motion capture file. Words such as “smooth” or “in sync” express an intention; they do not guarantee exact foot placement. If a particular choreographer's sequence matters, a filmed performance or an authorized motion editing workflow may fit better. This studio generates a new interpretation.

The most useful first question is therefore: what must remain clear? For a personal clip, that is usually the person's face and the recognizable side-to-side groove. Keep those priorities above background spectacle.

An original text-to-video prompt

Try this direction for a fictional person:

An adult dancer in a plain rust-colored shirt enjoys a relaxed Charme-inspired groove in a warmly lit open-air courtyard. Small side steps, soft knees and easy shoulder rolls. A few generic adults sway behind the lead. Keep the lead visible from knees upward in one steady vertical shot. Original instrumental soul atmosphere, without recognizable lyrics or a known recording.

This example establishes six useful constraints: an adult subject, unbranded clothing, a location, simple movement, clear framing and an original audio direction. It does not claim the subject is a real performer or that the courtyard is a particular Rio venue.

You can change the shirt or lighting without changing everything else. A controlled revision helps you understand what the prompt is doing. A completely different subject, setting, camera and dance makes comparison much harder.

An optional direction for a photo

With a photo, avoid rewriting the person's physical identity in text. The image already supplies visible facial features, hair and clothing. Conflicting descriptions can pull the output away from the reference.

A useful short direction is:

Keep the person and visible outfit from my photo. Place them in a warm evening courtyard, with relaxed side steps and gentle shoulders. Keep the camera near chest height and avoid a fast turn. Use a small background crowd that does not cross in front of the lead.

This asks for a new setting while preserving the most important reference information. It does not promise exact identity preservation, and it does not require you to upload a dance video. If your image is a close crop, the model still has to invent the unseen lower body. A clearer, wider input usually provides better guidance.

Plan a twelve-second scene in three beats

A short video benefits from fewer ideas. You can think of the opening, middle and ending without writing a film script.

These are creative directions rather than guaranteed timestamps. The model may interpret their timing differently. Do not advertise a precise second-by-second routine to somebody else before you have reviewed the finished file.

An uninterrupted shot also reduces the number of transitions that need to work. Cuts, zooms, wardrobe changes and crowd entrances each introduce another opportunity for visual inconsistency.

Describe the camera in everyday language

You do not need specialist terminology. “Keep the whole head and knees visible” is often more useful than an obscure lens specification. “The camera moves slowly” is clearer than a list of conflicting tracking effects.

For identity-led scenes, avoid demanding a full spin while also requiring the face to remain front-facing. For footwork-led scenes, a very tight portrait crop hides the movement you wanted to show. Choose a framing priority and keep the rest compatible with it.

A vertical composition suits a phone screen, but important details should remain away from the edges. Social platforms place interface elements around the frame. Review the actual upload preview before publishing rather than assuming the downloaded composition will look identical everywhere.

Keep the sound request original

Ask for an original instrumental mood if you want sound generated with the scene. Do not name a popular recording and assume that a generated video includes a license for it. This product does not provide the viral clip's original soundtrack.

If you later add music inside a social app, check that platform's current licensing rules for your account and use. TikTok explains that commercial content has different music considerations in its commercial music guidance. A song available for one personal post may not be suitable for an advertisement.

You can also mute the downloaded file in your editor. Generated audio is a creative output, so listen for unwanted speech or sounds before sharing it.

What a prompt cannot solve

A sentence cannot recover facial information that is absent from a blurry upload. It cannot make a crowded group photo unambiguous. It cannot guarantee that hands stay anatomically perfect during every movement.

Prompt changes are useful after input problems have been addressed. First inspect the source image for focus, lighting and clear ownership. Then inspect your instructions for contradictions. Only after those checks should you consider another paid render.

A completed video with an artistic imperfection is different from a failed task. Failed generation credits are returned. A new creative attempt is a new generation and consumes credits again. Keep a budget for experiments instead of treating unlimited retries as included.

A practical revision checklist

Before submitting, read the prompt once as if you were giving directions to a person filming a friend. Can they identify the lead, the movement, the place and the framing? Have you asked for several incompatible things? Is the scene safe and based on permission?

After generation, review the opening face, any turn, both hands and the separation between the lead and background adults. Watch with sound, then muted. A strong first frame does not establish quality across the whole clip.

Keep notes about changes that actually improved your result. Do not infer universal rules from one take. Generated videos vary, and a prompt that works for one photo may behave differently with another.

Questions about Charme dance AI prompts

Do I have to write a long prompt?

No. Image mode already supplies a dance direction. A short optional note is enough when you want a different mood or setting. A long prompt is useful only when its details work together.

Can I paste a competitor's full prompt?

You can describe the outcome you want in your own words. Copying a long prompt does not reproduce its model, reference inputs or production workflow. This guide provides original examples for this studio.

Can I request a celebrity or a child?

This studio is for fictional adults or consenting adult photos. Celebrity and public-figure photos are not accepted. See the photo guide before uploading.

Where do I try these directions?

Use the embedded studio below. Choose Text to video for a fictional adult or Image to video for a permitted photo, check the thirty-credit default cost, and prepare your scene. For the complete sequence, read how to make a Charme Dance AI video.