Verified against Runway Gen-4.5 · 2026-08-12
Generate a single high-contrast hero frame for a YouTube thumbnail from a video's key moment
A short, direct Runway Gen-4.5 still-frame brief for producing one punchy, legible hero image built specifically to survive being shrunk to a 120px-wide thumbnail.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Generate a single still frame, not a moving clip: a chef mid-flip with a pan of flames at the peak of the flame flare-up right as the pan flips, framed tight enough that the subject fills at least 60% of the frame. Expression and pose: exaggerated and legible — wide-eyed excitement, mouth open mid-shout — since this needs to read clearly at thumbnail size, not just at full resolution. Lighting: high-contrast, warm flame light on the subject against a darkened kitchen background, with a clear separation between subject and background so the silhouette stays readable even shrunk down. Background: simplified and slightly blurred version of a professional kitchen line, no competing focal points. Color: punchy, slightly boosted saturation on the orange-red flame, avoiding muddy mid-tones that flatten out at small sizes. Composition: leave clear negative space in the lower third of the frame for a text overlay to be added afterward — do not render any text in the generation itself. Do not generate: motion blur, multiple competing subjects, or fine detail that would disappear at small size.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
A thumbnail lives or dies at a size roughly 15 times smaller than the frame it was generated at, which is why this prompt front-loads instructions most video briefs skip entirely: subject fill percentage, exaggerated legible expression, and a named text-safe zone. Asking Gen-4.5 for a single still rather than a clip also matters mechanically — treating it as a video generation invites motion blur and mid-action ambiguity in the frame the model happens to land on, whereas explicitly requesting one still forces the model to commit to a single, deliberately composed peak moment instead of an arbitrary frame from a moving sequence. The high-contrast lighting and simplified, slightly blurred background instruction exists to preserve the silhouette read at small sizes — busy or evenly-lit backgrounds are the most common reason a technically good generation fails as a thumbnail once shrunk down, because the eye can no longer separate subject from background at 120 pixels wide. Explicitly telling the model not to render text keeps the composition clean for a real overlay added in an editor, since Gen-4.5's native text rendering is inconsistent enough that leaving it in-frame risks unreadable or misspelled pseudo-text that would need to be painted out anyway.
What you get back
A tight, high-contrast still of a chef mid-flip with an orange-red flame flare-up filling most of the frame, wide-eyed exaggerated expression, warm flame light against a blurred darkened kitchen line, clear open space in the lower third for a text overlay, no text rendered in the image.
Verified against
Runway Gen-4.5 Gen-4.5 · 2026-08-12
Changelog
- 2026-08-12 — Initial publish, 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
