The storm trend AI video is a short clip in which you walk into a silent crowd, stop in the middle and stay perfectly still while everyone around you breaks into the same dance at once. To make it, upload one photo to videosi, pick a crowd and a format, and download a 15-second silent video with no prompt to write. It is based on STORM II, the second part of a music video by GENER8ION with Yung Lean from April 2026. This guide answers the questions that come after the basics: the song, the original video, the meaning, Genjutsu, why faces drift, and what to check before you post.
Key takeaways
- The trend is a still person in the middle of a crowd that suddenly dances in unison, from STORM II by GENER8ION and Yung Lean.
- The original was filmed in Belgium and choreographed by Damien Jalet; videosi does not reuse any of that footage.
- We could not confirm any ban of "storm trend 2".
- You stay yourself in the middle of strangers, because faces drift when a crowd is large.
- The video is silent: add the sound from the platform library and label it as AI.
In this guide
- What is STORM II, and which part of the music video is the trend?
- What happens in the original storm dance video, and where was it filmed?
- What does the storm trend mean? What is the joke?
- Is Yung Lean in the storm AI trend?
- Was storm trend 2 banned?
- What is Genjutsu, and why do so many people use it for this trend?
- Is there a Higgsfield Genjutsu prompt I can copy?
- Why does the face not look like me when the crowd is big?
- Which crowd should I pick, and which photo works best?
- What does one storm trend video cost, and what if it fails?
- What should I check before I post it?
What is STORM II, and which part of the music video is the trend?
STORM is a double single and a two-part music video by GENER8ION, the duo of the French producer Surkin and the director Romain Gavras, starring the Swedish rapper Yung Lean. The Fader covered it on 24 April 2026. The first part, STORM, is the tense one. STORM II is the ecstatic second part, and it is the part people mean when they say "storm trend song".
The trend uses the moment in STORM II when suited schoolboys dance together on the school steps. That is why the AI versions are built around one long calm build and then a sudden, synchronized dance.
The clip you make on videosi is silent. You add STORM II yourself from the sound library of the app where you post, which also takes care of the music rights.
What happens in the original storm dance video, and where was it filmed?
In the original, Yung Lean plays a bully at an all-boys British boarding school. The dance in STORM II is the release after that tension: a large group of boys in school suits moves as one on the steps of the building. The choreography is by Damien Jalet.
The school is not in Britain. RTBF reports that it was filmed in Belgium, at the Collège Cardinal Mercier, dressed up as a British boarding school.
The videosi version does not use the original footage, cast or location. It recreates the structure with a generic crowd, and the first crowd option, dark school blazers, is the closest to the original look.
What does the storm trend mean? What is the joke?
The joke is contrast. One person stays completely still in the middle of a crowd, and then everyone around them starts moving in perfect unison. The still person is the only one who is not in on it, or the only one who refuses to be.
That is why the clip works for so many situations: the one sober friend at a party, the person who did not get the group chat message, the only adult who did not read the memo. The format needs a clear walk in, a long silent pause and one sharp drop, so the stillness feels uncomfortable before it becomes funny.
In the videosi clip the person never dances and never leaves the centre. If you wear something that looks different from the crowd, the joke reads faster.
Is Yung Lean in the storm AI trend?
Not in ours. Yung Lean is the star of the original music video, and the AI trend borrows only the idea of a still person and a dancing crowd. The videosi storm trend never shows his face, and it does not accept photos of celebrities or public figures.
Higgsfield Genjutsu, covered below, offers a Yung Lean preset so people can insert themselves into the STORM II clip. According to T-Journal, that preset lets you do this without writing a prompt. That is a different approach: it edits the original clip, while the videosi page builds a new 15-second scene from your photo.
Was storm trend 2 banned?
We could not confirm that. We found no official statement, platform notice or news report saying that a "storm trend 2" or any STORM II video was banned or removed, so we will not say it was.
What we can say is general. Platform rules change, and a post can be limited for reasons that have nothing to do with the trend, such as music rights or a missing AI label. If you want to lower the risk, use the platform's own sound library instead of a ripped audio file, and mark the video as AI-generated. Nobody can promise how any single post will be treated.
What is Genjutsu, and why do so many people use it for this trend?
Genjutsu is a video tool with two modes, Motion Transfer and Object Swap. T-Journal reports that it launched on 1 September 2026, and the storm trend spread soon after, because a ready-made Yung Lean preset made the clip easy to enter without any prompt writing.
That is a good reason for its popularity: a preset with one click is simple, and many of the early storm remakes were made that way. It is a legitimate tool and a sensible choice if you want to edit the original clip with yourself in it.
The videosi page is a separate, independent option. It starts from one photo and builds a fresh 15-second scene, with a choice of three crowds and three formats. You do not need an account anywhere else.
Is there a Higgsfield Genjutsu prompt I can copy?
The Genjutsu preset described above is built so that you do not write a prompt at all. You do not need one to use it, and we do not have an official one to point to.
On videosi the same applies: you choose a photo, a crowd and a format. The prompt is built in and we do not publish it. What we can describe is its structure. It has five parts: an identity lock for your face and outfit, a motion reference for the walk in and the drop, the crowd and place, a timeline with the crowd frozen until 11.5 seconds, and a list of rules such as no children, no text and no extra limbs.
If you try to write your own for a general video model, the usual failure is that the crowd starts dancing too early, or that every person gets your face. Those two problems are the reason for the motion reference and the identity lock.
Why does the face not look like me when the crowd is big?
There are two causes. First, your face is small. In most shots you are seen from behind or far away, and in the final high wide shot you are only a few pixels tall in a crowd of more than a hundred people. Second, models keep a face consistent far better with one or two people than with a whole crowd.
This is also why videosi does not try to turn every dancer into a copy of you. With that many people on screen, consistency drops and the copies stop looking like you. Instead we put you, in your own clothes, in the middle of strangers in a uniform. That keeps the one face that matters as clear as it can be.
Likeness is not guaranteed, and the last 2.5 seconds are where it drifts most.
Which crowd should I pick, and which photo works best?
There are three crowds:
- Dark school blazers on stone steps in front of an old brick building. This is the closest to the original music video.
- Nurses in scrubs in a hospital courtyard. It works well for jokes about work.
- Firefighters in dark gear on steps outside an old stone church. It is the most dramatic of the three.
For the photo, use one sharp, front-facing picture from the waist up in soft light, such as outdoors in the shade or by a window. Avoid sunglasses, masks, hats and flash. The outfit in the photo is the outfit in the video, so pick something that stands out against dark blazers or navy scrubs, like a light or bright colour.
Choose 9:16 for TikTok, Reels and Shorts, or 1:1 or 16:9 if you post elsewhere.
What does one storm trend video cost, and what if it fails?
A standard 15-second video (Standard model, 720p) costs 13 credits. There is no free generation and no free trial. Credit packs start at $6.99 for 10 credits and plans start at $29.90 a month. If a render fails, the credits come back automatically.
Rendering takes several minutes, and the result is saved in My creations if you close the page. The result is a silent MP4 without a watermark.
What should I check before I post it?
Run through this short list first:
- Add the sound in the app. The video is silent. Add the trend audio from TikTok's or Instagram's sound library, not from a downloaded file.
- Turn on the AI label. Mark the post as AI-generated where the platform offers it.
- Use your own photo, or one you have permission to use. Do not use another person's photo without their consent.
- No minors and no celebrities. videosi does not accept them, and you should not post them anywhere.
- Watch the whole clip once. Check the wide shot at the end for merged bodies, and if one looks wrong, generate again.
