What the prompt has to get right, and why you do not need to write it.
Updated Oct 29, 2026 · Independent fan-made tool. Not affiliated with, endorsed by or sponsored by K/DA, Riot Games or League of Legends. FIRESTARTER is a trademark and copyright of Riot Games.
Generator
Make your kda firestarter ai video
Three choices and one tap. The shot list, the timing and the prompt are already built.
The kda firestarter ai prompt
The kda firestarter ai prompt has to do four things at once: keep your face and outfit, follow a six-beat 13-second shot order, forbid anything not in your photo, and stay free of lip-sync and slow motion. On KDA Firestarter AI it is already written and tested. Make the video from one photo.
What the prompt has to solve
Keep the person identical to the photo, in the close-ups and the wide shots alike, without borrowing anything from the dancer in the reference.
Hold the timing: six beats in 13 seconds, from face close-up to arms-overhead finish, with the cuts landing where the sound does.
Set the rules that stop common errors: one adult, correct hands, no text or logos, no added stage costume, wings or wig.
Apply the chosen backdrop, a galaxy, silver smoke or a plain hallway, without changing the dancer.
Why copied prompts fail
It draws League of Legends characters or a generic K-pop stage. Asked for the dance by name, the model does not know the 13-second scene.
The hand at the lens is missing. Words alone rarely produce that beat or the cut from close-up to wide shot.
The face drifts. Close-ups and fast cuts are where likeness slips, and it slips further if the photo is dark or cropped.
A ponytail, wings or a costume appears. Models copy them from the reference unless told not to.
Each costs a paid attempt to find out, which is why pasting text rarely beats a template.
What is built in here
The motion template, the prompt for each of the three backdrops and the settings are fixed and tested on Seedance 2.0 / 2.5 by ByteDance, so every render follows the same shot list. You choose a photo, a backdrop and a format. The prompt itself is not published or editable. Likeness is not guaranteed, and a failed render returns its credits.
Clips from creators who made the effect by hand. Each video belongs to its creator and is shown here with permission and credit.
Three steps
How to do the kda firestarter ai trend
No timeline editing and no prompt writing. The clip is already choreographed, you bring the face.
01
Upload one photo
A clear, front-facing photo that shows your face, hair and the outfit you want to wear in the video.
02
Pick the template
Choose one of 3 backdrops: Galaxy Swirl, Silver Smoke or Real-life Cover. The dance is the same in all of them.
03
Download and post
Get the 13-second MP4 with the original sound already on it, and post it as it is.
Questions
FAQ
Is there a kda firestarter ai prompt I can copy and paste?
Prompts are passed around in tutorials, but a prompt alone does not make this video: it also needs a motion reference and the right inputs. On KDA Firestarter AI all of it is built in, so there is nothing to type.
Why does a copied prompt give a different result?
A copied kda firestarter ai prompt gives a different result because a prompt alone does not carry the motion: without the 13-second reference the hand at the lens and the close-up-to-wide cuts rarely appear. Models also drift from your face and add costumes. KDA Firestarter AI fixes the reference, the identity rules and the settings, so every render follows one shot list.
Which AI model does the prompt run on?
KDA Firestarter AI uses Seedance 2.0 / 2.5 by ByteDance.
