Ozil Goal AI Prompt
This is a free image-to-video prompt for the Ozil goal trend: one photo of a person in ordinary clothes goes in, and the prompt asks for the trend’s order of events at a night match filmed from the touchline. A keeper flies through the air, the ball pops up over the person’s head, they bounce it, a sliding defender goes down, the ball goes in. Set the length, the keeper’s kit and the ending below, then copy it.
Untested in other tools: it is an English paraphrase of the scene text our own maker runs, and we have not pasted it into any third-party tool. Treat it as a starting point you will adjust.
Vertical 9:16 video, 8 seconds, no dialogue, no subtitles, no on-screen text. Use the uploaded photo as the reference for the main person. Keep their face, face shape, skin tone, hair and identity the same for the whole clip; they are always the sharpest person in frame. Clothes: the person wears exactly the everyday clothes from the photo, unchanged, not a football kit. Opponents: two ordinary adult men with natural faces. The goalkeeper wears a plain green kit, the defender a plain white kit. Plain colour only: no lettering, no numbers, no stripes, no badges. Setting: a large football stadium at night, high floodlights making the turf bright, a full stand of blurred spectators, a row of glowing perimeter boards showing only moving bands of colour, with no letters, numbers or pictures. Camera: like a TV broadcast touchline telephoto, from the side and slightly above, the person in full length at about a third of the frame height, slight handheld shake, panning to follow them. 0 to 0.5s: the clip opens with the person already jogging with the ball. 0.5 to 2.5s: a goalkeeper launches a flying dive from the left: his whole body leaves the ground and flies flat through the air, arms stretched out. The person flicks the ball up with the toe and it pops high above their head; the keeper flies past their feet, lands on the turf and slides away. 2.5 to 3.5s: the ball drops; the person looks up and nods it once with the forehead, and it falls back to their feet. They keep dribbling. 3.5 to 5.5s: a defender slides in; the person cuts the ball inside, the defender slides past and ends up sitting on the grass. 5.5 to 8s: the person reaches the unguarded goal and pushes the ball into the net with the inside of the foot, then turns to the camera with a pleased grin while the crowd stands up and cheers. Audio: stadium crowd gasps and cheers and the sound of the ball being kicked. No commentary, no singing, no lyrics. Avoid: face swap, drifting features, a second sharp main person, the person in a football kit, extra fingers, extra legs, plastic skin, anime look, on-screen text, subtitles, scoreboards, shirt numbers, club badges, brand logos, watermarks, letters on the boards, the likeness of any real footballer, real kits, real brand logos, a keeper who only slides along the grass, a ball that never leaves the ground.
How to use the prompt
It is written for any image-to-video tool that accepts one reference photo plus a text prompt. We name no tool here because we have not run it in any.
- Pick the photo first. One adult, whole outfit visible if you can manage it, standing in decent light. The joke is the clothes, so a head-and-shoulders crop throws away half of it.
- Upload that photo as the reference image, or as the first frame if your tool only offers that. A first-frame upload tends to start the clip in the room the photo was taken in, which is why the setting line is written as a full description and not as “same place”.
- Set the tool to vertical 9:16 and to the same length you chose in the builder. If the tool’s length and the prompt’s timestamps disagree, the last beat is usually the one that gets cut, and the last beat is the goal.
- Paste the prompt whole. If the tool has a separate negative-prompt box, move the line that starts with “Avoid:” into it and delete the word “Avoid:”.
- Run it once before changing anything. Then use the fix list further down: change one line per run so you can tell what helped.
Rescaled from the trend’s clip
The five beats at 5, 8 and 10 seconds
The posts we pulled from the trend’s CapCut template run about five to six seconds each, and the order is always the same: flying keeper and the ball up first, then the bounce, then the defender, then the shot. The 8-second row below is the timing our own scene text uses. The other two rows fit the same five beats into a shorter or longer clip in the same proportions.
Longer is not automatically better. The flying keeper and the ball going up are what make someone scrolling past stop, and at ten seconds they take longer to arrive. If your tool offers five or eight seconds, start there.
| Clip | The run | The flying keeper | The bounce | The defender | The goal |
|---|---|---|---|---|---|
| 5 s | 0–0.3s | 0.3–1.6s | 1.6–2.2s | 2.2–3.4s | 3.4–5s |
| 8 s | 0–0.5s | 0.5–2.5s | 2.5–3.5s | 3.5–5.5s | 5.5–8s |
| 10 s | 0–0.6s | 0.6–3.1s | 3.1–4.4s | 4.4–6.9s | 6.9–10s |
Nine lines, nine reasons
Why each line is there
If you rewrite the prompt for your tool, these are the jobs each line is doing. Cut a line only if you know what you are giving up.
- Vertical 9:16 video… no on-screen text.
- Sets the shape before anything else. Generated captions and score graphics are a common way for invented words to appear, so the prompt rules them out at the top.
- Keep their face, face shape, skin tone, hair and identity…
- Ties the main person to the photo and keeps the photo’s face off the opponents. “Always the sharpest person in frame” stops a defender’s face from taking over.
- Clothes: exactly the everyday clothes from the photo…
- The line that carries the joke. A football pitch pulls hard toward football clothing, so the prompt names the clothes as something to keep instead of hoping the model leaves them alone.
- Setting: a large football stadium at night… moving bands of colour…
- Night, full stands, floodlit turf and perimeter boards are what make the frame read as a televised match. Asking for colour bands instead of words keeps real brand names off the boards.
- Opponents: two ordinary adult men… plain colour only…
- Two opponents is the minimum the trend needs: the keeper who flies past and the defender who ends up sitting. Plain kits with no lettering avoid a real club’s shirt. In the trend posts the keeper often wears 1; the prompt leaves numbers out because generated numbers drift.
- Camera: like a TV broadcast touchline telephoto…
- The trend posts look filmed side-on by a broadcast long lens, the person about a third of the frame tall. A camera behind the goal or overhead gives you a different video, however good the moves are.
- The five timed lines
- Timestamps stop a model from spending the whole clip on the run. The keeper line spells out “whole body leaves the ground”, because a keeper sliding along the grass is the most common way this beat comes out wrong.
- Audio: stadium crowd gasps and cheers… No commentary…
- Generated commentary tends to invent names. Leave music out too and add a sound from the platform’s own library when you post.
- Avoid: …the likeness of any real footballer, real kits, real brand logos.
- Keeps the clip your own: your face, an invented match. It also lowers the chance of a tool refusing the request.
Change one line per run
When the clip comes back wrong
These are the failures this kind of shot is prone to, and the edit we would try first. They are reasoned from how the prompt is built, not from a log of our own renders.
| What you see | Try this |
|---|---|
| The face drifts into someone else | Use a sharper photo with the face larger and unobstructed. Drop to the five-second length. If it still drifts, remove the defender beat so there is less time and fewer bodies on screen. |
| The outfit is swapped for a football kit | Describe the clothes in the character line in plain words (“a red cardigan over floral pyjamas”) and choose an opponents’ kit colour that is nothing like them. A full-body photo helps more than any wording. |
| The keeper slides along the grass instead of flying | Keep “his whole body leaves the ground and flies flat through the air” and the Avoid entry “a keeper who only slides along the grass”. Models prefer the easier motion, so say what the body does, not what the move is called. |
| The ball never goes up | Make the keeper line blunter about the ball: “the ball leaves the ground and passes high above the person’s head”. Without that moment the clip stops reading as this trend. |
| Six legs in the slide tackle | Limbs tangle when two bodies overlap. Ask for the defender to “slide past in front of the person without touching them”, and re-roll: this one varies a lot run to run. |
| Readable brand names on the boards | Keep “made-up nonsense lettering only” and add “blurred by motion” to the setting line. If a real name still shows, do not post that take. |
| The clip ends before the ball is in the net | The tool’s clip length is shorter than the prompt’s. Match them, or switch the ending to “cut on the shot”, which needs less time than the turn to the camera. |
| The camera cuts or jumps behind the goal | Add “one continuous shot from the touchline, no cuts” to the end of the camera line. The trend posts do cut around the shot at goal, but a cut early on is where a face most often changes. |
| The request is refused | Check you have not added a real player’s or club’s name. Check the photo is of an adult, clothed, and not a public figure. |
What this prompt cannot promise
It paraphrases the scene text our own maker runs, and the clip beside this text was made from that scene text on this site, not from this English version in another tool. We have not run this English version anywhere, so there is no success rate to quote for it.
A text prompt also cannot fix a weak photo. If the face is small, turned away or behind sunglasses, no wording will make the runner look like that person for several seconds of sprinting.
Each attempt in a third-party tool costs whatever that tool charges, and physical action with three bodies and a ball is the kind of shot that often needs several tries. Budget for re-rolls.
Use your own photo or an adult who has said yes. Do not put a real footballer, a celebrity or a child into it, and label the post as AI-generated where the platform asks.
This page is an independent guide. It is not affiliated with the footballer the trend is named after, or with any club, federation or broadcaster, and the prompt asks for no real person’s likeness.
Or skip the prompt
The maker on our home page has a scene for this trend built in: you add one photo and press the button, with no prompt to write. A video is 10 credits, from $9.90, 8 seconds at 9:16. There is a clip from it beside the limits above. We have only a couple of renders of the current scene, so how often a run comes out clean is not measured yet; a render that fails on our side returns its credits.
Asked in the comments
Ozil Goal AI Prompt: questions
What is the Ozil AI prompt?
A text prompt for an image-to-video tool that places the person in your photo, in their own clothes, into the trend’s football clip: a goalkeeper flying past through the air, the ball popping up over their head and bounced off the head, a sliding defender left on the grass, and the ball into the net at a night match filmed from the touchline. The builder on this page writes it for three clip lengths.
Which AI do I paste it into?
Any image-to-video tool that takes one reference photo and a text prompt, and can output vertical 9:16 at five seconds or more. We have not tested this English version in any third-party tool, so we do not recommend one over another.
Is the prompt free?
Yes. Copy it, change it, use it anywhere. What costs money is rendering in whichever tool you paste it into. The maker on this site is the paid shortcut, from $9.90 a video.
Why does my person end up in a football shirt?
Because the scene is a football match and the model fills in what it expects to see there. Keep the line about the exact clothes from the photo, describe the outfit in words, use a full-body photo, and pick an opponents’ kit colour that differs from what the person is wearing.
Should I write the footballer’s name in the prompt?
No. The prompt works by describing the moves, and naming a real player invites the model to draw his face or a real shirt, which is exactly what you do not want and what many tools refuse. The Avoid line asks for no real footballer’s likeness on purpose.
Five, eight or ten seconds?
The trend’s posts run about five to six seconds; eight is what our own scene uses. Ten gives the moves more room but opens more slowly and gives the face more time to drift. Start with five or eight.
Can I use the CapCut template instead?
Yes, and for many people that is the simpler route: the posts we pulled were all made with one CapCut template that takes one photo. A prompt is for people who want to change the length, the kit colour or the ending, or who use a different tool. The CapCut page on this site compares the routes.