About the Raindance AI Video Generator

What the Raindance AI video generator does

The trend has a very specific look: a weathered wooden pier, a calm sea, low golden-hour sun and a 4:3 frame, with two people singing to each other. Getting that look from a text prompt takes many tries, and the faces rarely stay the same from one take to the next. We took a different route. We make the clip first, get it right once, and then put you into it. You upload a photo of each person, choose a template, and download an 8-second MP4 that you finish in TikTok or CapCut by adding the song.

The goal is simple: the clip should look like the posts you see in the trend, and it should look like you. Everything on the site, from the template design to the upload tips, serves those two things. Open today: Pier at Sunset and The Orange Room. More templates open after they pass our checks. A Solo clip costs 15 credits and a Duo clip 30 credits. Prices are listed in full on the pricing page.

How a Raindance AI video template is made

Every template starts from the shots that keep coming back in the most-viewed posts of the trend. We write down what they have in common (the pier, the tyres along its side, the direction of the light, how close the camera is, how slowly it moves) and design a scene around those details. We then generate a still frame with an image model, animate it with a video model, and pick the take that reads as the trend straight away. Each clip is cropped to 4:3 and its audio is replaced with quiet ambient sound, so no song or generated singing is baked in.

The people in every template, and in the examples on this site, are AI-generated. We do not use photos of real people taken from the internet, and we never use or generate the likeness of Dave, Tems or any other celebrity. Before a template goes live we run it with several sets of test faces and redo it if the swap shows visible glitches.

How we pick which templates to make

We only make a template when the same shot keeps coming back in the most-viewed posts of the trend, because people recognise it in the first second. Then every template has to pass the same rules before it goes live:

• Plain, well-fitting clothes with closed necklines, no layers and no prints, because the swap can invent patterns on large plain areas of fabric.
• Nothing in the frame that carries text: no signs, logos, screens or phones, and hands kept free.
• Two sets of test faces, checked frame by frame, especially the last two seconds and the hands. If either set shows a glitch, the template is remade, not released.

That is why there are only a few templates at any time. We would rather open one clip that works every time than ten that work sometimes.

How the swap works

When you press Generate, our video engine looks at the first frame of the template and finds each person in it. It then follows that person through the whole clip and replaces them with the person in your photo. Before that, the tool crops your photo to your face and hair, which keeps the swap clean: your face and hair come from your photo, the scene and the moves follow the template, and the engine fills in the outfit. Everything else stays as it was: the pier, the sea, the light, the camera move and the timing of the gestures. That is why a clear, front-facing photo gives the best result.

In Duo mode the two people are swapped one after the other, the person on the left first and then the person on the right, which is why a Duo clip takes about three minutes and a Solo clip about a minute and a half. If a generation fails, the credits are returned automatically.

Your photos and privacy

• Photos are used only to make the video you asked for. They are deleted within 7 days and are never used to train models.
• Photos and your videos are never shown in our gallery or anywhere else on this site.
• Generated videos are kept for up to 30 days so you can download them again, then deleted.
• You can ask us to delete your account and everything in it at any time.

The details are in our privacy policy.

What you can and cannot make

• Only upload photos of yourself or of people who agreed to it. Duo mode asks you to confirm this before generating.
• No photos of celebrities, public figures or anyone under 18.
• No sexual, hateful or harassing content.
• No videos meant to deceive, such as passing a clip off as real footage of someone.

We refuse generations that break these rules and act on removal requests within 24 hours. The full rules are in our terms of service.

How we write our guides

• Guides are written by the Raindance AI Editorial Team, and every step is tested on our own tool before it is published.
• Numbers about the trend, such as views and dates, come from the posts themselves and say when we checked them.
• Each guide shows the date of its last update. When we find a mistake we correct it and update the date.
• We link to the sources we used, such as Google Trends and public artist pages.

Independence and how we make money

Raindance AI is not affiliated with, endorsed by or sponsored by Dave, Tems, their labels, TikTok or CapCut. “Raindance” is used only to describe the trend. We earn money by selling credits for the tool. We do not run ads and we do not take payment for anything we mention in our guides.

Who builds it

Raindance AI is made by a small independent team that builds single-purpose AI tools: one job per site, done well, with prices and limits written in plain numbers. We follow the trend closely and add templates when new versions of the Raindance look start getting views.

Contact

Email support@raindanceaivideo.com. A person reads every message and we reply within one business day. For billing, privacy and removal requests, see the contact page for what to include.