La Tormenta de Arena AI is a 10-second nightclub clip in which a person leans on a balcony, stops smiling and stares down at the dance floor while Dorian's song "La tormenta de arena" plays. videosi makes the solo version from one photo of you, with no prompt, and the file is silent so you add the song yourself.
The trend page covers what it is and how to make it. This guide goes deeper on the questions people ask afterwards: where the name comes from, what the film connection really is, why Spanish speakers found it first, and how to pick a photo and a starting point in the song.
Key takeaways
- The song is "La tormenta de arena" by the Spanish band Dorian, released in 2009; the line "te he perdido entre la gente" means "I have lost you in the crowd".
- Tutorials link the balcony scene to the film Tres metros sobre el cielo, but the film's Wikipedia plot does not describe a nightclub balcony, so treat that link as what creators say, not as a verified source.
- "Mind Enterprises balcony" is a search phrase some people use; we found no Mind Enterprises track in these videos.
- videosi makes you alone on the balcony and an anonymous crowd below, never a second real face.
- A standard 10-second video is 9 credits, and the video is silent.
In this guide
- Why is the trend tied to "te he perdido entre la gente"?
- What is the song and where does it come from?
- Is the balcony scene really from Tres metros sobre el cielo?
- Why did it spread in Spanish first?
- What is the difference between a CapCut template and an AI video?
- Why do people search for Mind Enterprises balcony?
- What makes a good photo for the balcony stare?
- Where in the song should the video start?
- Can I use a pet or a friend's photo?
- Will a post like this get taken down?
- What does it cost?
Why is the trend tied to "te he perdido entre la gente"?
Because the line explains what the person on the balcony is doing. "Te he perdido entre la gente" translates as "I have lost you in the crowd". The face that is smiling in the first shot goes quiet in the second, and the stare is a search: somebody down there was in sight a moment ago and is now gone in the dancing.
That is why creators time the lyric to the close-up. The words arrive as the eyes lock onto the floor, and the viewer fills in the rest of the story without any caption. It is also why many people name the trend after the lyric rather than the song title, and why the alias "te he perdido entre la gente" brings up the same videos as La Tormenta de Arena AI.
What is the song and where does it come from?
According to Wikipedia's article on Dorian, the band is a Spanish pop rock group from Barcelona that mixes new wave, electronic and indie rock. "La tormenta de arena" is a 2009 song from their third album, La ciudad subterránea. It was released as a single and got a music video the same year, the first in a three-part video series for that album. The band later recorded an acoustic version for the 2015 unplugged album Diez años y un día.
Is the balcony scene really from Tres metros sobre el cielo?
This is what the tutorials say, and it is why you will see the names "3MSC balcony scene" and "Tres metros club stare". We could not verify it. Wikipedia's page on the film, released in Spain on 3 December 2010 and directed by Fernando González Molina, summarises a teenage romance between Babi and H, a street racer, set against a house party that ends in a fight and a birthday party. The summary does not describe a nightclub or a balcony.
That does not mean the claim is wrong. A plot summary leaves out many scenes, and a short moment can become the image a trend is built on. It does mean we do not state it as fact. On the page we describe 3MSC as the scene tutorials point to, and the earliest AI versions we found are dated 2 October 2026 with no single original post identified.
Why did it spread in Spanish first?
Three things line up. The song is in Spanish, the lyric is the hook, and the tutorials that took off from mid-October 2026 were titled in Spanish, for example a "how to make the video where you look from a balcony" style of search. Our trend radar first flagged the topic on 16 October and marked it as spreading on 24 October.
A language-bound lyric travels along language lines at first. Spanish speakers recognise the song and feel the line, so the clip lands without explanation. Viewers who do not speak Spanish still read the visual story, which is why the trend then appears under English names such as "balcony stare" and "club balcony".
What is the difference between a CapCut template and an AI video?
A CapCut template is a prebuilt edit. You drop your photos into slots, and the template moves, crops and times them to the music. The film-scene templates that circulated in October 2026 paste two photos into the scene: one person on the balcony and one on the floor. The result looks like the original because it is a cut of existing footage.
An AI video is generated from scratch. videosi sends your photo plus a short motion reference to a video model and renders a new 10-second scene, so the face is yours in every shot, including the ones from behind. The trade-off is that likeness is not guaranteed and each render costs credits, while a template is quick to apply and free to repeat.
They are not rivals. Many people render the clip on the La Tormenta de Arena AI page, then open CapCut only to add the song, trim the ending and export for TikTok or Reels.
Why do people search for Mind Enterprises balcony?
Some searches pair the trend with "Mind Enterprises balcony", and we cannot say exactly why. Wikipedia describes Mind Enterprises as an Italian duo of Andrea Tirone and Roberto Conigliaro whose music mixes Italo disco, techno and 1990s style electro, with a debut single, "Summer War", in 2012. The same article says they became the subject of viral memes around 2025.
In the balcony stare videos we checked, we found no Mind Enterprises track. The sound was Dorian's. One guess is that searchers remember a different balcony meme, or a DJ set, and merge the two. That is only a guess, and we would rather say we do not know than invent a link.
What makes a good photo for the balcony stare?
The clip spends about half its length close to your face and half behind your head, so the photo has to settle both. Pick one where your hair is clearly visible, your face points at the camera, the light is even and your shoulders are in the frame. The scene is dark with amber behind you, so dark or neutral clothing holds its shape better than neon.
A calm face is better than a big laugh. The video goes from a smile to a serious gaze, and an open mouthed laugh is hard to turn into the stare. Hats, hoods and heavy filters make the shots from behind drift, since the model has to guess the back of your head.
Where in the song should the video start?
The clip is 10 seconds and the key moment is the close-up stare at roughly 5 to 6.5 seconds. Place the song so that "te he perdido entre la gente" lands there. In practice that means starting the audio a little before the line, so the smile and the lean play over the build-up and the stare arrives on the words.
We have not mapped the exact second for every version of the track, so listen once and drag the audio until the line meets the close-up. The cut to the over-the-shoulder shot at about 6.5 seconds then falls on the next phrase, and the spotlight at the end gives you a natural place to stop.
Can I use a pet or a friend's photo?
A friend, yes, if they agree. Use an adult who has said yes to being in the video, and post it only with their consent. The friend is the person on the balcony; videosi does not paste a second real person onto the dance floor, and it does not make two person face swaps or intimate scenes.
The scene was written for a person, with hands on a rail and a shift of expression, so a photo of a pet is not what it was tuned for and we have not tested it. If you want an animal balcony version, expect to experiment. Our page links related ideas, but we do not promise a pet result.
Will a post like this get taken down?
We cannot speak for any platform, and rules change. What we can say is that this clip is a stylised, cinematic scene of one adult, with no second identifiable person, no text and no claim that it is real footage. Even so, label it: TikTok, Instagram and YouTube ask creators to mark realistic AI generated content, and ticking that option is the safest habit.
Do not use someone else's photo without permission, and do not present the video as a recording of a real night out. If a platform removes a post, the usual cause is a missing label or a complaint from the person shown, which is why consent matters.
What does it cost?
A standard 10-second video (Standard model, 720p) is 9 credits. Credits come in packs, and the Single pack is US$6.99 for 10 credits. The price is shown before you press Generate, and faster or higher resolution tiers cost more.
That makes one finished video roughly the price of one pack, so it is worth choosing the photo with care before you start. When you are ready, open the La Tormenta de Arena AI generator, upload one photo and pick 9:16, 1:1 or 16:9.
