The K/DA Firestarter AI dance is a TikTok trend built on one 13-second vogue scene from the FIRESTARTER music video, and you can make your own version from one photo. This guide answers the questions that come after the basics on the trend page: what the vogue moves are, why cats took over, how to post it safely and how to pick a photo and a backdrop.
Key takeaways
- The trend is a short vogue scene from K/DA's FIRESTARTER video, released in October 2026 and animated by Fortiche.
- The scene runs in six beats: face close-up, walk-in, hand at the lens, vogue and kick, arms frame the face, arms-overhead finish.
- FIRESTARTER is copyrighted by Riot Games: post on TikTok, Reels or Shorts under the platform's music licence, never in ads.
- You dance as yourself in your own outfit; no League of Legends character is drawn.
- A 13-second Standard 720p video is 11 credits; a first video costs 10, so the $6.99 Single pack covers it.
In this guide
- What is the vogue scene everyone is remaking?
- Who is in K/DA now, and who made the video?
- What is Kai'Sa actually doing in the vogue scene?
- What are the six beats, second by second?
- Why did the cat versions blow up?
- What is the difference between a real dance cover and an AI version?
- How do you post the song without trouble?
- How do you shoot the photo?
- Which of the three backdrops should you pick?
- Why does the dance fail in a general video tool?
- What does it cost to try?
What is the vogue scene everyone is remaking?
It is a roughly 13-second moment from the animated FIRESTARTER video, where Kai'Sa dances alone in a swirling purple galaxy. It opens on a face close-up, ends on a pose with both arms crossed overhead, and in between has a hand thrust at the camera that bursts into light.
The trend started days after the video came out. From about 20 October 2026, AI creators put other dancers into the same scene, and people filmed real covers at home or in hallways. It has only just started spreading, so this guide quotes no view counts and names no first creator. The K/DA Firestarter AI dance page turns the scene into a clip of you.
Who is in K/DA now, and who made the video?
K/DA is the virtual group from League of Legends. According to The Korea Herald, FIRESTARTER is its first release in six years, with Miyeon and Soyeon of i-dle joined by Sophie Powers and Mimi Webb, who make their K/DA debuts. The four champions are Ahri, Akali, Evelynn and Kai'Sa.
Dexerto reports that Fortiche, the studio behind Arcane and the original Pop/Stars video, co-produced the new one, and that fans called it worth the wait. Outlets differ by a day on the release date (18 or 19 October 2026, depending on time zone), so treat it as the weekend of 18 to 19 October.
What is Kai'Sa actually doing in the vogue scene?
Vogue here means the stylised, angular arm and hand posing of vogue dance, not a casual freestyle. The arms do the work: they sweep, frame the face, cross and rise, while the body stays confident and the eyes stay on the camera.
Three details make it read as Firestarter. First, the hand thrust at the lens, where the palm pushes toward the viewer and light fills the frame. Second, the arms framing the face before a kick back. Third, the final cross overhead with glowing hands. In the AI version the glow is a visual effect added by the model, not something you have to act out. Keep the mouth relaxed; the scene is about posture, not lip-sync.
What are the six beats, second by second?
This is the structure used on the trend page, taken from a fan-made remake of the scene, not from the official video, so poses are close but not frame-exact.
- 0 to 1 s, face close-up. Tight on the face in pink haze, looking straight into the camera.
- 1 to 3.5 s, walk-in. A wider shot: step forward through the smoke, one hand at the chest, then sweep the arm out.
- 3.5 to 6 s, hand at the lens. A low-angle close shot, then the hand thrust that bursts into light.
- 6 to 9 s, vogue and kick. Wide shot: vogue arms with the light ribbons, a high kick and a turn.
- 9 to 11 s, arms frame the face. Medium shot, both arms around the face, one leg kicks back, arms rise.
- 11 to 13 s, final pose. Arms crossed overhead, eyes on the camera.
If you want to learn the real-life version, practise the hand thrust and the kick separately, then link them with the arm sweep. Film full body so the kick stays in frame.
Why did the cat versions blow up?
A cat doing a confident vogue is funny because the contrast is so large: a fluffy animal with a dead-serious face, moving to a glossy pop-star scene. It also needs no explanation, so it travels fast across feeds, and a cat has no face to be recognised by, which makes it easy to share.
The same logic explains kittens, a cat in a bob wig and other characters in the first wave of remakes. The joke only works while the scene is recognisable, which is why it spread within days of the video. Many viral remakes star cats, but videosi's version is tuned for one adult person, so pets may come out less reliably.
What is the difference between a real dance cover and an AI version?
A real cover is you filming yourself, usually at home or in a hallway, often with the music video playing on top of the screen. It shows your own skill and its mistakes are part of the charm. It needs space, practice and a few takes.
An AI version needs none of that: one photo goes in and a 13-second clip comes out, with the poses and camera moves of the scene. The trade-offs are that hands can blur for a few frames during the light burst and that likeness is not guaranteed. The third backdrop on the trend page is a plain hallway look, so the AI clip can pass for an ordinary phone-filmed cover. Say it is AI when you post; TikTok asks creators to label realistic AI content.
How do you post the song without trouble?
The original song, FIRESTARTER, is copyrighted by Riot Games. The videosi clip comes with the trend's original sound already on it, and that is meant for posting on TikTok, Reels or Shorts, where the platform's own music licence covers it.
- Post on the platform. Upload the MP4 as it is, and let the platform handle the sound licence.
- Do not use it in ads. The platform licence does not cover paid promotion or brand content.
- Do not re-upload the song. Do not add your own copy of the track on top of the clip.
videosi is independent and not affiliated with K/DA, Riot Games or League of Legends, and it does not generate League of Legends character likenesses. You appear as yourself in your own outfit.
How do you shoot the photo?
Stand, face the camera and show your whole body from head to shoes in even, bright light. The outfit in the photo is the outfit you dance in, and the scene has a high kick and wide shots, so legs and shoes in frame keep the outfit right when the camera pulls back.
- Keep the face sharp. The clip opens on a face close-up, so skip sunglasses, masks and heavy beauty filters.
- Wear something with shape. A cropped top, wide trousers or a jacket reads well in the glow; very dark, shapeless clothes can melt into the galaxy.
- Avoid dark or blurry selfies. They tend to drift during the cuts and the light burst.
Use only your own photo or one from an adult who agreed. Photos of minors, celebrities and public figures are not accepted, and users must be 18 or older.
Which of the three backdrops should you pick?
There are three looks, and the right one depends on where you post.
- Purple galaxy swirl. The closest to the music video: violet stars, pink smoke and magenta light ribbons. Pick it for the full trend look.
- Silver smoke and light ribbons. A dark void with cool, pale ribbons. Pick it if purple clashes with your outfit or you want a moodier clip.
- Real-life dance cover. A bright school hallway with no glow. Pick it for the cover look, where the light burst becomes a quick hand toward the lens.
Then choose a format: 9:16 for TikTok and Reels, or 1:1 or 16:9.
Why does the dance fail in a general video tool?
You can try, but four things tend to go wrong. Ask a model for the K/DA Firestarter dance by name and it draws the League of Legends characters or a generic K-pop stage, because it does not know the 13-second scene.
- No hand thrust. Without a motion reference, the hand at the lens and the cut from close-up to wide shot rarely appear.
- Drifting likeness. Close-ups and fast cuts are where the face slips.
- Surprise costumes. Models add stage costumes, wings or a ponytail from the reference unless the prompt forbids anything not in your photo.
The fix is a motion reference, an identity lock and explicit rules, which are built into the trend page.
What does it cost to try?
There is no free generation. A 13-second video on the Standard model at 720p is 11 credits, and a first video on an account costs no more than a 12-second one, which is 10 credits, so the $6.99 Single pack of 10 credits covers it. Plans start at $29.90 a month if you make many, and a failed, timed-out or blocked render returns its credits automatically.
To make yours, open the K/DA Firestarter AI dance page, upload one photo, pick a backdrop and press Generate.
