Brand mascot prompts that keep the mascot on-model
A mascot only works if people recognize it instantly, which means the same proportions, the same colors and the same few gestures every time. AI video models drift toward a generic cute character unless the references hold them back. These prompts cover greetings and signature moves, scenes with your product, and seasonal campaigns, each written so the mascot's design comes from the images and the prompt only directs the action.
What makes mascot prompts work
Build the reference set like a model sheet: front, three-quarter, side and back views on a plain background, in the final render style, whether 3D or flat 2D, plus a close-up of the face. That is five images, the most RefVid takes per subject. If the mascot wears a logo, make sure it is visible in one of them.
If the mascot has no mouth or unusual proportions, show that clearly; models tend to add mouths and shrink heads.
Give the mascot a small vocabulary of gestures and reuse it: a two-handed wave, a hop, a head tilt. Repeated signature moves make a mascot feel like a character, and simple moves keep the body on-model.
Keep brand colors in the references, not the prompt. Writing "bright orange" when the references are a specific orange often pushes the shade off.
The examples below call the mascot Pip; use the name you saved yours under, and RefVid passes it to the model as the mascot's reference. Clip length and aspect ratio are settings, shown here in brackets.
Greetings and signature moves
Prompt (5 s, 9:16): Pip stands in front of a bright shop entrance, hops once on the spot and waves with both hands toward the camera. Static camera at Pip's eye level, full body in frame. Soft daylight, shop window blurred behind. Same 3D render style as the references.
Why it works: eye-level framing keeps proportions true, and a hop plus wave is a clear, repeatable greeting. Use this as the opening of every video in a series.
Prompt (4 s, 9:16): Pip pops up from the bottom edge of the frame against a plain yellow background, gives a thumbs-up and tilts its head to one side. Static camera, medium shot. Even studio light, soft shadow.
Why it works: a pop-in with a plain background is easy to reuse as an end card. If the hand changes shape, swap one of the references for a close-up of the mascot's hand.
Prompt (6 s, 16:9): Pip stands on a small round stage and takes a slow bow, then straightens up and spreads its arms wide. A spotlight from above, dark background, light confetti falling. Camera slowly pushes in.
Why it works: a bow is a single smooth arc the model handles well. Confetti adds celebration without any extra character.
Mascot with your product
Save the product as its own subject with its own photos, then name both in the prompt. Up to three subjects can share one shot. The examples use two invented products saved as MOVA box and CRUNCHO box.
Prompt (7 s, 16:9): Pip walks along a sidewalk carrying the MOVA box in both arms, then sets it down at a front door and gives a little wave. The camera tracks beside it at Pip's height. Sunny afternoon, suburban street.
Why it works: carrying a box in two arms is simpler than holding a small item in one hand. Keep the box large relative to the mascot so its branding stays readable.
Prompt (5 s, 16:9): Pip sits next to the CRUNCHO box on a breakfast table, points at the box with one hand and nods. Bright morning kitchen light, a bowl and spoon nearby. Static camera, medium shot.
Why it works: pointing keeps the mascot and product separate in the frame, which avoids the two designs blending.
Seasonal and campaign scenes
Prompt (6 s, 9:16): Pip catches snowflakes on an outstretched hand in a quiet park at dusk, then looks up at the camera. Warm lamp light, snow falling slowly, trees with fairy lights behind. Camera static at Pip's height.
Why it works: weather does the work of the season without costume changes. If you add a scarf or hat, save a separate seasonal version of the mascot with its own references.
Prompt (7 s, 16:9): Pip runs across warm sand toward the sea, stops at the waterline and splashes its feet as a small wave arrives. The camera follows from behind, then holds. Bright midday sun, turquoise water.
Why it works: a run that ends in a simple action gives energy with a clean ending. Back-view references help, since much of the clip shows the mascot from behind.
Running a mascot campaign in RefVid
RefVid stores the mascot as a character and each product separately, so a campaign of weekly clips reuses the same model sheet every time. A scene photo, such as your shop front, can set the location, and a free still of the scene shows the framing before you pay for a clip.
For a short story, one clip can hold up to three shots, and "next shot" continues from the last frame of the previous clip, so a greeting, a product moment and an end card can follow each other in the same setting.
Frequently asked questions
- Why does my mascot get a mouth or bigger eyes?
- AI video models drift toward generic cute character features. Show the mascot's face clearly from several angles and avoid prompts that ask it to talk or smile with its mouth.
- Can a mascot speak in AI video?
- Yes, if its design has a mouth: with sound on, the model can generate a short spoken line written in quotation marks. For a mascot without a mouth, keep sound off, let it gesture, and add a voiceover in your editing app.
- Should mascot references be 3D renders or drawings?
- Whichever style the mascot videos should have. Use one style for the whole reference set; mixing renders and sketches confuses the output.
- How many gestures should a mascot have?
- Three to five signature moves is plenty. Reusing the same gestures builds recognition and keeps the mascot's body stable in every clip.