Ozil AI Prompt: Copy, Customize and Use

What prompt should I use for an Ozil goal AI video?

Describe who should appear, what supplies the movement and which parts of the scene matter. For motion-control tools, attach both the person’s photo and an appropriate reference clip. A prompt cannot replace a missing reference or guarantee a faithful face.

A starter prompt for a general video tool

Create a fictional football highlight featuring the person shown in the attached photo. Keep their facial features and clothing consistent. The attacker approaches a rushing goalkeeper, lifts the ball over the keeper, controls it beyond the challenge and finishes into the open goal. Use a steady wide broadcast view that keeps the attacker and ball visible. Show natural body movement and a continuous, plausible ball path. Do not add captions or logos.

Use this starting point in a tool with a custom prompt field. Different models accept different inputs and durations. If your tool supports motion references, assign your clip to its motion input and follow that tool’s instructions. Use only reference media you are entitled to upload.

How do I use this prompt?

1. Copy the prompt above into your video tool’s prompt field.
2. Upload your person photo in the image or character reference field.
3. If supported, upload the full goal sequence in the motion reference field.
4. Choose the duration and aspect ratio in the tool’s controls; writing “15 seconds” in a prompt does not change a model’s duration limit.
5. Generate once, watch the result, then change one instruction or input for the next attempt.

A shorter prompt for motion-reference tools

Use the attached person photo for the character’s appearance and the attached video for movement. Preserve the visible outfit and face throughout the run, chip and finish. Keep the player and ball in frame with a steady wide camera. No captions or added logos.

Use this version when the motion input already supplies the action. For a wider view, add “Keep the full body, goalkeeper and goal visible.” For distracting camera cuts, add “Use one continuous shot.” These instructions guide the result; output still depends on the tool and inputs.

Do I need to paste a prompt on ozilai.online?

You do not need to write a custom prompt for Ozil AI. The generator pairs your photo with a prepared football motion reference. Choose your output quality and review the credit cost before generating. Follow the step-by-step guide to prepare your first edit.

Why does the result look like a different person?

A small, blurry or partly hidden face gives the model less usable detail. Start by changing the photo, not by adding increasingly long identity instructions. Use one person, even lighting and a visible face. Motion models can still change identity during a turn or fast movement; inspect several points in the clip.

Why is the chip or final shot missing?

Check whether the reference actually contains the full action and whether the model’s duration limit fits it. A generic image-to-video model may invent a different sequence even when the wording is precise. Avoid squeezing several separate actions into a very short generation.

How should I improve a bad result?

Change one input at a time and keep notes: photo, reference, prompt or duration. Compare the same moments in each output. If the face is wrong, improve the image; if the action is wrong, inspect the reference and model mode. Do not keep paying for unchanged retries expecting a guaranteed outcome.

See the photo requirements and guide index. Technical source: Kie Wan 3.0 reference input specifications.