How to make a
Raindance AI video.

Start with two portraits and a favorite moment. The studio supplies the reference and built-in prompt; each generation creates one 10-second part.

1. Pick a section of the reference

The 30-second sequence has three parts: Part 1 (0:00–0:10), Part 2 (0:10–0:20), and Part 3 (0:20–0:30). Choose one and watch its preview. Your selection determines the actual clip used for generation.

2. Upload a male photo and a female photo

Use a JPG, PNG or WebP under 10 MB and at least 300 × 300 pixels. Show one face and upper body in even light. Avoid group photos, sunglasses, heavy filters and faces that fill only a tiny part of the image.

3. Check who plays each role

The male upload replaces the original male performer; the female upload replaces the original female performer. These roles remain fixed across cuts and screen-position changes. A solo close-up stays a solo close-up: the prompt does not add the other person. Use Swap uploaded photos only if you placed the images in the wrong slots.

4. Choose quality and keep the original framing

Every output is 10 seconds. Choose 768P for 160 credits or 2K for 260 credits per part. Outputs use 16:9 with padding around the source frame. Cropping, stretching and new background content are excluded from the prompt to preserve the reference composition. See credit packs and three-part costs.

5. Generate, preview and download

When the service opens, sign in, add credits, confirm permission to use both photos and generate. Keep the selected part visible in the summary before submitting. My videos keeps your part labels, status and MP4 downloads. A confirmed failed job restores its reserved credits.

Make a full 30-second video from three parts

1. Generate Part 1, Part 2 and Part 3 separately. Reuse the same male and female portraits and quality. Do not swap roles when the performers change positions.
2. Download all three MP4 files. Add them to CapCut or another editor in Part 1 → Part 2 → Part 3 order.
3. Review the cuts. Separate AI generations may differ in faces, clothing or movement; trim or transition as needed. Automatic stitching and seamless continuity are not included.
4. Export the assembled video. The three full clips total 30 seconds before any edits. They cost 480 credits at 768P or 780 at 2K.

Add the Raindance sound when posting

The reference preview includes source audio, but generated audio may differ. This tool does not grant music rights. If the track is available in your region and for your account, add it through TikTok or Instagram’s licensed music library. Follow the platform’s AI disclosure rules and use only photos you have permission to use.

If the result does not match your photos

Try clearer portraits with unobstructed faces, confirm the selected part and check the role assignment. AI can still change appearance, gestures and timing. A new attempt is a separate paid generation; exact likeness is not guaranteed.

The 30-second reference and all three parts are ready to preview. Paid generation remains closed until the generation service is connected.