Skip to content
AI Raindance Video

Quick actions

Appearance

Two adult performers in Halloween costumes dancing together in rain

TWO PEOPLE. ONE SCENE. NO SWAPPED FACES.

Halloween Duet Video

A duet is not two solo videos standing next to each other. It is one scene with two roles. This workflow asks for two separate adult photos, maps each photo to a role and keeps the scene's camera and movement consistent. That is the difference between a couple clip and a collage.

Input
2 adult photos
Length
8-10 seconds
Format
Silent MP4

AI PHOTO LOOK

Try the look on your photo

Upload an adult photo and preview an original costume portrait. One look uses 20 credits.

Your credits0
This look20 credits

Choose a look

Photos stay private. Only upload adults who have consented.

Image routing uses KIE with fal, Replicate and WaveSpeed fallbacks where supported; moderation uses SEEAPI.

What makes a duet scene different

A duet scene has two performers, a shared camera and a relationship between the bodies. The model has to keep both identities, both costumes and the choreography coherent. That is harder than a solo clip, and it is why role mapping matters.

The scene preview shows the original performers. Your photos replace them. The preview is not a promised result; it is a reference for the camera angle, movement and duration. Watch it before you upload so you know what the model is trying to preserve.

The two-photo rule

Use one photo per adult. Do not use a single group photo, a screenshot of a couple or a photo where one face is hidden. The model needs a clean image for each role. Similar framing and lighting make the result more coherent.

If you are using a couple photo from a wedding or party, crop each person into a separate image first. Make sure both adults have consented to the edit. The upload is private, but privacy is not the same as permission.

  • Photo 1: the person on the left or the leading role.
  • Photo 2: the person on the right or the turning role.
  • Both faces visible, both photos sharp.
  • Similar light and distance are better than matching outfits.
  • No children, no non-consenting adults, no public figures.

Costumes that work in a duet

The best duet costumes have contrast without visual clutter. Cowboy and Angel works because the silhouettes are distinct. Gothic and celestial works because the colors separate. Two skeleton costumes can work if the scene is close and the makeup is the focus.

Avoid two costumes with large wings, long trains and props. The model has enough to track with two bodies. One hero element per person is the practical limit.

How to direct the scene

Keep the direction short. 'Keep both faces recognizable, preserve the costumes and follow the scene's movement' is enough. If you add a story, the model may prioritize the story over the identity. The scene already carries the choreography.

If one person's face drifts, check the source photo before changing the prompt. A turned face, a shadow or a low-resolution image will cause more problems than a missing adjective.

After the clip is ready

The download is silent. Add a track you have the right to use, trim the clip if needed and publish. Label AI-generated content where the platform requires it. If you are making a couple announcement, make sure both people are comfortable with the final image.

A duet clip is also a good test for a longer sequence. Generate two short scenes with the same photos and join them in an editor. The result feels more like a story and less like a single loop.

Pricing and limits

A duet video uses 100 credits, the same as a solo video. The photos are uploaded separately and the scene handles the two roles. Failed jobs return credits. The result is a short, silent MP4, not a full music video.

If you only need a couple portrait, use the couple photo editor first. It costs less and lets you judge the costumes before you animate. If the portrait is strong, the duet video is the next step.

Related tools

Make a video

Upload two separate adult photos, choose a duet scene and generate a short silent clip.

Create your Raindance video

The scene and movement stay. You choose who steps in.

Add your photos

JPG, PNG or WebP · up to 10 MB

Use a clear photo of one adult facing the camera. For full appearance, include the outfit. In duets, photo 1 replaces the first role and photo 2 the second.

Choose your scene
What should change?
Extra direction (optional)

A public result may be chosen for our gallery after review. Your uploaded photos stay private.

Preview scenes for free. Create with credits.

Scene previews are free to browse. Creating your own video uses credits from a paid pack. Previously issued bonus credits remain valid until their original expiry. New welcome and daily credits are no longer offered.

$9.99/ 1 video

100 credits · 8 seconds

One-time payment. Failed generations return your credits.

Your videos

Choose a scene above. Finished videos will appear in this space.

Questions people ask

Can I use one photo for both people?

No. Use two separate photos, one per adult. A single group photo is not a reliable input for a two-person role mapping.

Can the two people wear different costumes?

Yes. Contrast is good. Keep each costume readable and avoid too many props or large wings.

Does the video include music?

No. The output is a silent MP4. Add a track you have the right to use in the app where you publish.

How much does a duet cost?

A duet video uses 100 credits, the same as a solo video. A still image look uses 20 credits.

What if the faces get swapped?

Use two clean photos with similar framing, name the left and right roles, and keep the costumes distinct. A separated pose is easier than overlapping faces.

Can I make a duet with a pet?

The duet scenes are designed for two human performers. A pet is better handled as a separate pet portrait.

Can I use the clip commercially?

You are responsible for the source photos, likeness rights and music. Both adults must consent to the use.

How long is the clip?

Prepared duet scenes are 8-10 seconds. The original 30-second trend is split into separate clips.

Updated: