The Stormzy Shut Up AI trend puts someone else at the front of the park crowd in Stormzy's 2015 "Shut Up" freestyle, rapping "tell my man, shut up" to the camera. Stormzy Shut Up AI makes that 10-second clip from one photo of you, with the bars already on the video, and the 10 seconds it uses contain no swear words.
This guide goes deeper than the FAQ on the trend page. It covers why that stretch is clean, how to choose between the two scenes, why the mouth is not exact, how the AI clip differs from the 2015 video, which photo works and what a render costs.
Key takeaways
- The 10 seconds run from "man try say he's better than me" to the last "shut up", and that stretch has no swearing, although other parts of the full freestyle are explicit.
- You choose between a seated crew on park benches and a standing park crowd; both are 10 seconds and both put you at the front.
- Your mouth moves with the rhythm of the bars but is not synced word for word.
- A 10-second Standard 720p video is 10 credits, there is no free trial, and a failed render returns its credits.
- Label the clip as AI-generated and use a photo you have the right to use.
In this guide
- Why does this 10-second clip have no swearing?
- How is the AI clip different from the 2015 video?
- Seated crew or standing crowd: which scene should I pick?
- What happens second by second?
- Why doesn't the mouth match every word?
- How should I shoot the photo?
- Why do the faces in the crowd look slightly different?
- Why can't I just type the trend into a video model?
- What does one clip cost, and what if it fails?
- Can I post it, and do I need to label it?
Why does this 10-second clip have no swearing?
Because the trend only uses one short stretch of the freestyle. The clip starts at "man try say he's better than me", runs through "tell my man, shut up", "mention my name in your tweets" and "oi, rudeboy, shut up", and ends on a last "shut up". None of those lines contains a swear word.
The full 2015 freestyle is a different matter, and parts of it are explicit. That is why the choice of window matters. It is also why people on TikTok and Reels keep reusing exactly these seconds: the hook is the most recognisable part, and it is easy to post without bleeping anything. If you edit the clip further, for example by adding a longer sound from the original track, listen to the whole thing first, because you may bring in lines that are not on the clip we render.
How is the AI clip different from the 2015 video?
The original is a freestyle that Stormzy put on YouTube in May 2015. The Fader's write-up from 18 May 2015 describes it as a new freestyle over the early grime beat "Functions On The Low", with fast, intense rhymes and a casual video: Stormzy in a park with friends who back him up and sometimes sing backing vocals.
According to Wikipedia's entry on the song, the studio version was released in the UK in September 2015, and after a Christmas number one campaign launched on 13 December 2015 it climbed to number 8 on the UK Singles Chart. That story is the reason the park clip has such a long life online.
The AI version is not the 2015 footage. The motion references we use are fan-made AI park freestyle clips, so the crowd, the park and the framing look like the trend clips rather than the original video. The performer at the front is replaced by you, and the sound that is on the clip is the freestyle audio for those 10 seconds. To be in the shot yourself, use the Stormzy Shut Up AI page.
Seated crew or standing crowd: which scene should I pick?
There are two scenes, and the difference is mostly about mood and framing.
The seated crew is the default. You stand at the front while a group sits on park benches behind you, laughing and nodding. It reads as the classic "mates watching the freestyle" shot, and it is the closest to what most people remember. It works best with a photo of you in a bold top, because your outfit stands out against the seated group.
The standing park crowd puts you at the front of a larger group of people on their feet. The crowd is denser and the energy is higher, so it suits a bigger, louder look. In that scene you also finish with an arm thrown up, which gives the clip a clear last beat.
If you are unsure, pick the seated crew for a first try. Both cost the same, both are 10 seconds, and you can choose 16:9, 9:16 or 1:1 in either. If the format is going to TikTok, Reels or Shorts, 9:16 is the usual choice.
What happens second by second?
The clip has five beats. From 0 to 2.6 seconds you stand at the front of the crew, grinning and rolling your hands as "man try say he's better than me" lands, then "tell my man, shut up".
From 2.6 to 4.7 seconds there is a sharp "shut up" and a flick of the hand, then you point at the camera on "mention my name in your tweets". From 4.7 to 6.9 seconds your hands spread wide toward the lens on "oi, rudeboy, shut up" while the crew behind you laughs.
From 6.9 to 8.4 seconds you pat your chest on "better than me?" and lean in with a grin. From 8.4 to 10 seconds comes one last "shut up", and you hold the grin to the camera as the clip ends.
It is one continuous shot with no cuts, which helps. The face never turns far from the lens, so the likeness has a steadier job than in clips that jump between angles.
Why doesn't the mouth match every word?
The video model receives a motion reference and a short instruction to swap the rapper for the person in your photo. It copies timing, posture and the rhythm of the mouth, but it does not read the bars and place each syllable. The sound is added after the video is made, so the two are lined up by timing, not generated together.
For a fast freestyle, most viewers read this as performance rather than error, especially because your hands, grin and chest pat carry so much of the clip. If you need frame-accurate lip-sync, you would need a dedicated lip-sync workflow on a longer talking shot. For this trend, the attitude matters more than the syllables.
How should I shoot the photo?
Use one clear photo of you, face to the camera, head to waist, in even light. The clip is a waist-up shot with your hands moving the whole time, so a head-to-waist photo gives the model your real build and clothes. A tight selfie leaves it to invent the rest.
Keep sunglasses, masks and heavy filters out of the picture. You rap straight into the lens, so a front-facing face holds the likeness best. Both scenes are filmed under a grey British sky, so a photo in soft daylight or shade fits better than harsh sun or a flash.
Your clothes come from the photo. A red track jacket is the classic nod to the original, but any bold colour reads well against the darker crowd. Use one person per photo, because the clip shows a single rapper at the front.
Why do the faces in the crowd look slightly different?
Only the rapper is replaced. Everyone behind you stays as the people in the reference clip, and because they are small and moving, their faces may shift slightly from frame to frame. That is a known limit of swapping one person inside a busy shot.
It also means you cannot place a friend into the crowd. The clip has one swapped performer, and the rest of the group is fixed.
Why can't I just type the trend into a video model?
You can try, but the result is usually a generic rapper in a park. The crowd packed behind the performer and the timing of the bars only come from a real motion reference, which is why this page pairs your photo with a 10-second park freestyle clip.
Two other problems come up. A prompt that says "rap like the man in the video" tends to keep the original rapper instead of swapping him out, so the swap and the moves have to be described separately. And video models often refuse a reference built on a famous person's face, which is why the references here are non-celebrity park freestyle clips. We do not publish the full prompt, but the page shows its structure: who is replaced, what leaves with them, the motion reference, the performance line and the likeness rule.
What does one clip cost, and what if it fails?
A 10-second Stormzy Shut Up AI video on the Standard model at 720p costs 10 credits. A Single pack is $8.90 for 10 credits, so that is one clip, and plans start at $29.90 a month. Better models and resolutions cost more credits.
There is no free generation and no free trial. If a render fails, times out or is blocked, the credits are returned to your account automatically. A result you simply do not like is not refunded, so spend your effort on the photo first. A good photo is the cheapest way to avoid paying for a second try.
Can I post it, and do I need to label it?
Use your own photo, or the photo of an adult who has said yes. Do not upload someone's picture without asking. The clip is a fan-made style video and is not affiliated with Stormzy or his label, and it is not footage of a real appearance.
Label the video as AI-generated when you post, using the platform's AI label if it has one, so nobody mistakes it for the real thing. You can keep the sound that is on the clip or pick the trend sound on the app you post to. If you add a commercial track, follow that platform's music rules, because we make no claim about licensing on your account.
If you want the shortest route, make it once from a good photo, check the sound and post. The page has the two scenes and the formats ready.



