Verified against Kling 3.0 · 2026-08-12
Generate a talking clip that keeps the lower third clear for caption overlays
A Kling 3.0 brief that composes the shot with caption placement in mind from the start, so the finished clip doesn't need to be re-framed or have its subject cropped awkwardly to make room for burned-in captions later.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Subject and action: a man in his 30s in a plain t-shirt, seated talking directly to camera about why most people over-water their houseplants, counting off points on his fingers kept above chest height so hand movement doesn't drop into the lower frame. Framing: subject's face and shoulders framed in the upper two-thirds of the vertical frame; the bottom third of frame stays visually simple — plain background, no text, signage, or busy detail — since that space will carry burned-in captions after generation. Camera: static or a very slow, minimal push-in only, so caption timing doesn't have to chase a moving subject. Environment and lighting: a plain painted wall in a home office, even front-facing light so the face reads clearly at small sizes once captions are added below it. Style and mood: clear, direct, presenter-style delivery, unhurried pacing with natural pauses between sentences to give captions room to display one clean phrase at a time. Audio: single clear voice, no background music competing with speech clarity, consistent volume throughout.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
Captions are almost always added after generation, so the framing decision has to be made now, at prompt time, rather than fixed later — a clip generated with gestures or background detail filling the lower third forces an editor to either shrink and crop the subject to make room for text or place captions over the speaker's hands, both of which look amateurish. Explicitly keeping gestures above chest height and specifying the bottom third stay visually plain gives Kling 3.0 a concrete compositional target instead of a vague "leave room for text," which the model has no way to interpret consistently on its own. Slowing the camera to static or minimal push-in and asking for natural pauses between sentences isn't just a delivery style choice — caption software times text to speech cadence, and a subject who talks in a continuous rush without pauses produces long, cramped caption lines, while natural sentence breaks let each caption chunk land as one clean, readable phrase.
What you get back
A static medium shot: a man in a plain t-shirt against a simple wall, upper two-thirds of frame, counting points off on his fingers at chest height while explaining over-watering, clear even lighting, unhurried pacing with a natural pause between each point, bottom third of frame left clean for captions.
Verified against
Kling 3.0 3.0 · 2026-08-12
Changelog
- 2026-08-12 — Initial publish, verified against Kling 3.0.
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
