Verified against Runway Gen-4.5 · 2026-07-25

Keep the same character consistent across a multi-shot brand clip

A Runway Gen-4.5 brief structured around a locked character reference so a mascot or recurring presenter looks like the same person or character shot to shot, not a new face each generation.

Runway Gen-4.5

The prompt

Ready to copy — highlighted parts are example details you can swap.

Character reference: a friendly orange fox mascot in a blue zip hoodie with rounded, soft features, kept visually identical to the reference image in face, outfit and proportions across every shot.
Subject and action: the character waving and gesturing toward a product display in a a bright, minimal retail storefront.
Camera: static locked-off medium shot, no camera movement, so attention stays on the character's consistency rather than the frame.
Environment and lighting: soft even daylight from large front windows, consistent color temperature across shots to support later editing continuity.
Style and mood: flat, friendly 3D-toy aesthetic matching the existing brand illustrations, clean and on-brand, no stylistic drift from shot to shot.
Audio: ambient (quiet retail store ambience) only, no dialogue, reserved for a voiceover track added separately.

Customize the highlighted detailsoptional — the prompt above already works

The character's fixed visual details, matched to a reference image

What the character is doing in this specific shot

Where the shot is set

The lighting description, kept identical across the shot series

The brand's visual style to hold across every clip

The background ambient sound only · optional

Why this works

Gen-4.5's actual differentiator versus earlier Runway models is reference-image-driven character consistency across separate generations, so this prompt is built to be re-run shot after shot with the same character reference and the same fixed description clause held word-for-word — only the action/location/lighting variables change between shots. Locking the camera to a static shot removes one more variable that otherwise competes with the model's attention for consistency budget; a moving camera plus a character it needs to keep matching is a harder joint problem than either alone. Keeping dialogue out of the generated audio and treating ambient sound as the only audio layer is deliberate: it keeps voice/lip-sync (a much weaker point for consistency across shots) out of the clip entirely, so a separate voiceover can be added in post without needing to match generated mouth movement.

What you get back

A static medium shot of the same orange fox mascot from the reference image, hoodie and proportions unchanged, waving toward a product display under soft window light, quiet store ambience underneath, no camera movement and no stylistic drift from the previous shot in the series.

Verified against

Runway Gen-4.5 Gen-4.5 · 2026-07-25

Changelog

2026-07-25 — Published, verified against Runway Gen-4.5.

Need this built into your business?

If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.

EXPLORE WHAT SCULT BUILDS