How detailed should the prompt be?

One sentence for the subject, one for the dance style, one for the setting, one for the camera, and one for the output format. The studio accepts prompts up to 2,000 characters, but a focused 60-100 word prompt is usually more reliable than a long list.

Is text to video cheaper than image to video?

No. For the same duration and settings the credit cost is the same. The difference is the input, not the price. Check the cost line in the studio before submitting.

What is text to dance video best for?

Testing a concept, making a stylized character clip, or producing a background loop when you do not have a photo. For a recognizable family member, image-to-video is the better starting point.

Text to Dance Video AI — Describe the Clip You Want

Q: Do I need a photo for text to dance video?

No. Text-to-video starts from the written scene and movement, so you can test an idea before you have a usable photo. If you later want a specific child or character, switch to image-to-video and upload an authorized photo.

Q: Can I choose the dance style?

Yes. Name a style the video model recognizes: K-pop, hip-hop, ballet, salsa, disco, or a simple TikTok loop. Asking for several styles in one clip usually produces an unclear result.

Quick answer

Text to dance video creates a short dance clip from a written description. You do not need a photo or a reference video. You write the subject, the dance style, the movement, the setting, and the camera, and the studio sends that prompt to a hosted video model. The result is a 5-second or 10-second clip you can download.

This mode is best for testing an idea, making a stylized character clip, or producing a loop when you do not have a usable photo. If the clip needs to feature a specific child or family member, use photo to dance video instead.

What should a text-to-dance prompt contain?

A reliable prompt has five parts. Keep each part to one sentence.

Part	What to write	Example
Subject	Who is dancing and one identifying detail	A cheerful toddler in a mint-green tutu dress
Style	One dance style	A gentle K-pop bounce
Movement	Intensity and one action	Slow, controlled arm movements; no jump
Setting	One room or scene, one lighting note	A soft pastel studio with warm lights
Camera	Format and one camera move	Vertical 9:16 medium shot, slow push-in

A full example:

A cheerful toddler in a mint-green tutu dress performs a gentle K-pop bounce in a soft pastel studio with warm lights. Slow, controlled arm movements, no jump. Vertical 9:16 medium shot, subject centered, slow push-in. Keep the face and body proportions consistent, one subject only, natural hands and feet, no text, no logos, no watermark.

The prompt is written in English because hosted video models generally follow English instructions most consistently. You can write your notes in any language and translate the final prompt.

What does it cost?

Text to video uses the same credit pricing as image to video:

Duration	No audio	With audio
5 seconds	300 credits	600 credits
10 seconds	600 credits	1,200 credits

The studio prints the cost of the current setting before you submit. There are no free credits; the Try pack is $9.90 one-time for 600 credits, enough for two standard 5-second clips. The pricing page lists every pack and plan.

Text to dance video vs. image to dance video

	Text to dance video	Image to dance video
Input	A written prompt	A photo + a prompt
Subject	Generated by the model	The person in the photo
Best for	Concept tests, stylized characters, loops	A specific child, a family member, a real portrait
Identity	No real identity to preserve	The model tries to preserve the photo's subject
Cost	Same per duration	Same per duration

Use text to video to answer "does this costume and setting read well in five seconds?" Once the idea works, switch to image to video if you want a real child in the clip.

Five prompt mistakes that produce a bad clip

Stacking styles. "K-pop, ballet, and hip-hop" produces an unclear blend. Pick one.
Asking for a story. A 5-second clip cannot show a beginning, a middle, and an end. Describe one moment.
Too many people. One subject is the reliable default. Group scenes increase face-swap and limb errors.
No camera instruction. Without a framing sentence, the model may crop the subject or drift wide.
No guardrails. Add "one subject only, natural hands and feet, no text, no logos, no watermark" to reduce common failures.

If the first draft is not usable, change one part of the prompt and generate again. Changing five parts at once makes it impossible to learn what helped. The prompt generator builds the five parts for you, and the Gemini and ChatGPT guide shows how to draft and refine the wording in a chat tool.

What can you do with a text-generated clip?

Test a holiday idea before buying a costume. See the Halloween guide and Christmas guide.
Make a stylized character loop for a landing page or a background video.
Prototype a dance challenge before filming a real version with a child.
Create a vertical clip for TikTok, Reels, or Shorts. Keep the subject centered and add music in the platform editor.

Text-generated clips are also useful when privacy matters: if you do not want a real child's face in the model, describe a character instead of uploading a family photo.