Kling's "Subject → Movement → Scene → Camera → Lighting" Prompt Formula, Rebuilt for Veo3Gen: 10 Copy-Paste Templates for Ads & Reels

TL;DR

Kling’s official prompt structure—Subject → Movement → Scene → Camera → Lighting—works because it forces you to describe a filmable action and how it’s shot, not just “a vibe.” Kling publishes this as Subject (description) + Movement + Scene (description) + optional Camera Language + Lighting + Atmosphere (https://kling.ai/quickstart/text-to-video-prompt-guide).

This post compresses that into a reusable 5‑slot prompt card for Veo3Gen, plus 10 copy‑paste templates for ads & reels and an iteration method (lock 4 slots, vary 1) so you can reach a usable clip faster—and keep a campaign consistent.

Key takeaways

• Write prompts like directions to a scene, not a shopping list of objects—this tends to improve adherence and clarity. (https://blog.fal.ai/kling-3-0-prompting-guide/)
• Keep your prompt readable: 2–4 sentences or labeled lines often produce more stable, detailed results than one-line prompts. (https://picsart.com/blog/kling-ai-prompts/)
• When a clip fails, fix in this order: Movement → Camera → Scene (most “weird” outputs start with unclear action).
• Iterate with an A/B grid: lock 4 slots, vary 1 so you learn what actually changed the output.
• In Veo3Gen, use Lite for cheapest previews, Fast as a great default, and Quality for maximum fidelity once the prompt is proven.

Why standardize on a 5-slot prompt formula

Creators don’t usually fail because they “can’t prompt.” They fail because the brief is missing one of these:

1. What is the main subject? (Kling defines “subject” as the main focus—person/animal/object/etc.) (https://kling.ai/quickstart/text-to-video-prompt-guide)
2. What changes over time? (Kling advises movement should be straightforward and suitable for a short clip.) (https://kling.ai/quickstart/text-to-video-prompt-guide)
3. Where is it happening? (Kling’s “scene” includes the environment and foreground/background elements.) (https://kling.ai/quickstart/text-to-video-prompt-guide)
4. How is it filmed? (Camera language can be expressed in natural directions like close-up, wide shot, push-in, pan, tilt, tracking.) (https://kling.ai/blog/kling-ai-prompt-guide)
5. What does it feel like? (Lighting + mood)

Standardizing gives you repeatability: teammates can swap one slot (e.g., Scene) without accidentally changing the message (Movement).

The original Kling structure (what we’re rebuilding)

Kling’s quickstart formula is:

• Subject (Subject Description)
• Subject Movement
• Scene (Scene Description)
• Optional Camera Language
• Lighting
• Atmosphere

(https://kling.ai/quickstart/text-to-video-prompt-guide)

Two practical adjustments for ad/reel workflows:

1) Merge Atmosphere into Lighting

Instead of splitting “Lighting” vs “Atmosphere,” treat “atmosphere/mood” as a modifier inside the Lighting line (e.g., “soft window light, cozy mood”). Fewer fields = more usage.

2) Make it readable on purpose

Kling-focused guidance repeatedly points toward “director-style” prompting:

• Write it like directions to a scene, not a list. (https://blog.fal.ai/kling-3-0-prompting-guide/)
• Use 2–4 sentences for stability/detail rather than one-line prompts. (https://picsart.com/blog/kling-ai-prompts/)

The Veo3Gen 5‑Slot Prompt Card (copy‑paste)

Paste this into your notes and reuse it for every ad variant.

Subject: [WHO/WHAT is the focus? Include 1–2 identifiers that must stay consistent.]
Movement: [ONE clear action that changes over time; simple enough for a short clip.]
Scene: [Where it happens; add 2–3 anchors: foreground + background.]
Camera Language: [Shot size + one camera move (or locked-off).]
Lighting: [Direction + softness + time-of-day + mood.]

Prompt hygiene rules (keep these strict)

• One action per clip. If you wrote “opens, applies, reacts,” you wrote three clips.
• One camera move max. (Or none: “locked-off tripod.”)
• No contradictions. Don’t mix “handheld” with “perfectly locked-off.”

Worked example (before → after) you can reuse

This is the exact upgrade pattern that saves rerolls.

Before (vague one-liner)

“Make a cinematic video of a woman using a vitamin C serum in a bathroom, nice lighting, trendy.”

Why it rerolls forever:

• “Using” isn’t specific (open? drop? apply? reveal?).
• “Cinematic” doesn’t tell the model what the camera does.
• The bathroom can drift because there are no anchors.

After (campaign-ready 5-slot card)

Subject: Woman (late 20s) holding a dropper bottle of [BRAND] vitamin C serum, label facing camera, clean nails.
Movement: She squeezes one drop onto her fingertip, dots it once on her cheek, then turns her face slightly to reveal a visible glow.
Scene: Minimal bathroom vanity, round mirror edge visible, folded towel, small plant in the background.
Camera Language: Tight medium close-up, slow push-in, steady.
Lighting: Soft diffused side window light, warm morning mood, gentle highlights on skin.

What changed (so you can copy the pattern)

Slot	Before	After	Why it helps
Movement	“using”	“one drop → dot → turn to reveal”	Adds a clear mini-story and end state
Scene	“bathroom”	vanity + mirror edge + towel + plant	Stabilizes the environment
Camera	“cinematic”	“tight medium close-up, slow push-in”	One explicit move; less chaos
Lighting	“nice lighting”	direction + softness + mood	Anchors the vibe without buzzwords

10 copy‑paste templates (ads, reels, product, creator hooks)

Each template is already in the 5-slot format. Replace bracketed items.

1) DTC skincare texture reveal (macro)

Subject: A clean glass jar of [BRAND] hydrating gel moisturizer, label facing camera; a hand with neat nails.
Movement: The hand scoops a small amount and smears it on the back of the other hand to show glossy texture.
Scene: Minimal bathroom vanity, white marble counter, folded towel in background.
Camera Language: Extreme close-up macro, slow push-in, shallow depth of field.
Lighting: Soft diffused bathroom light, slightly warm, highlights on gel.

2) UGC “stop scrolling” hook (authentic phone feel)

Subject: Relatable creator holding [PRODUCT], casual hoodie, natural face.
Movement: They step into frame, point at the product, and mouth: “Wait—don’t buy this until you see…”
Scene: Bedroom/home office, slightly messy but real, posters on wall.
Camera Language: Front-facing phone camera feel, chest-up medium close-up, slight handheld sway.
Lighting: Daylight from side window, soft shadows, realistic.

3) App feature tap-and-reveal (over-the-shoulder)

Subject: Hand holding a smartphone showing [APP NAME] with a clear UI.
Movement: Thumb taps [FEATURE BUTTON] then scrolls to reveal [RESULT].
Scene: Coffee shop table, latte art, notebook, street bokeh in background.
Camera Language: Over-the-shoulder close-up, gentle tilt down following the thumb.
Lighting: Bright café daylight, crisp but not harsh.

4) Food hero pour (thirst trap)

Subject: Tall glass with ice and [DRINK], condensation visible.
Movement: Slow pour from bottle; liquid cascades over ice; bubbles rise.
Scene: Modern kitchen counter, citrus slices, cutting board.
Camera Language: Tight close-up, locked-off tripod.
Lighting: Strong side light like late afternoon sun, sparkling highlights.

5) Beauty “apply then reveal” (single continuous action)

Subject: Creator holding [PRODUCT], hair pinned back, clean skin.
Movement: One smooth application stroke, then a small head turn to show finish.
Scene: Neutral backdrop, vanity mirror edge visible, makeup items out of focus.
Camera Language: Medium close-up, gentle pan following the hand.
Lighting: Soft ring-light look with clean catchlights, even skin tone.

6) Local service proof (wipe-to-shine)

Subject: Professional cleaner in branded shirt [BUSINESS NAME], spray bottle and microfiber cloth.
Movement: One pass wipe on dusty glass reveals a clear shine expanding across the frame.
Scene: Sunlit living room, large window, indoor plants; dust visible in light beam.
Camera Language: Close-up on glass, small pull-back as clean area grows.
Lighting: Bright natural sunlight beam, high-contrast sparkle.

7) Ecommerce unboxing (top-down)

Subject: Shipping box with [BRAND] tape; hands with simple rings.
Movement: Slice tape, open lid, lift product into frame and hold for one beat.
Scene: Wooden desk, packing paper, branded insert card partially visible.
Camera Language: Top-down overhead shot, steady, slight push-in as lid opens.
Lighting: Soft studio light, neutral white, minimal shadows.

8) Fitness micro-demo (one rep)

Subject: Athlete wearing [BRAND] resistance band around legs, focused expression.
Movement: One slow controlled squat rep; band visibly stretches.
Scene: Simple gym corner with mat; mirror blurred in background.
Camera Language: Low-angle medium shot, locked-off.
Lighting: Cool gym lighting, clean and bright.

9) SaaS “mess → order” metaphor (no UI required)

Subject: Stressed office worker at laptop; sticky notes labeled [TASKS].
Movement: They sweep sticky notes into a neat stack, exhale, then smile at the cleared desk.
Scene: Modern desk, second monitor with charts blurred, coffee mug, daylight window.
Camera Language: Medium shot, slow push-in at the moment the desk clears.
Lighting: Neutral daylight, slightly desaturated, calm mood.

10) Fashion/accessory hero (confidence + texture)

Subject: Model wearing [ACCESSORY] (watch/bag/sunglasses), confident posture.
Movement: Adjust the accessory once, then turn slightly so the material catches light.
Scene: Urban sidewalk, clean storefront reflections, subtle background motion.
Camera Language: Medium close-up, gentle tracking shot alongside subject.
Lighting: Golden hour side light, warm highlights, soft background.

Iterate fast: the lock‑4 / vary‑1 A/B grid

This is the simplest method that teaches you what your prompt is doing.

1. Lock Subject + Movement + Scene + Lighting. Vary only Camera Language.
2. Pick the best camera. Now lock everything and vary Lighting.
3. Only then vary Scene (keep Movement unchanged).

Mini table: symptoms → smallest fix

Symptom	Likely cause	Fix slot first	Smallest effective change
Looks static	No time-based change	Movement	Add one reveal: open/pour/wipe/tap/turn
Action unclear	Too many verbs	Movement	Reduce to one continuous action
Dizzy/chaotic	Too many camera moves	Camera	Choose one move or locked-off
Wrong vibe	Lighting underspecified	Lighting	Add direction + softness + time-of-day
Random background	Scene too vague	Scene	Add 2 anchors: foreground + background

Where Veo3Gen fits (without changing your formula)

Once your prompt card works, Veo3Gen lets you run the same structure across deliverables:

• Text-to-video and image-to-video.
• First-and-last-frame control on Veo 3.1 for tighter continuity.
• Native, synchronized audio (dialogue/SFX/music) generated in a single pass.
• Output options: 720p, 1080p, 4K (4K on Veo 3.1 Fast/Quality), aspect ratios 16:9 and 9:16.
• Three modes: Veo 3.1 Lite (cheapest previews), Veo 3.1 Fast (quick, great default), Veo 3.1 Quality (max fidelity).
• Pricing model: pay-as-you-go credits + optional monthly plans, and purchased credits do not expire. New users get free credits to start.

Mid‑article CTA: If you’re ready to turn these templates into repeatable variants (hooks, angles, 9:16 vs 16:9) without rebuilding prompts from scratch, use Veo3Gen’s Fast mode for iteration and Quality for final selects—then scale the same card across your campaign.

Checklist

• Use the 5 labeled slots (Subject, Movement, Scene, Camera, Lighting)
• Keep Movement to one filmable action with a visible end state
• Add 2–3 Scene anchors (foreground + background)
• Specify shot size and one camera move (or locked-off)
• Specify lighting direction + softness + time-of-day
• Keep prompts readable: labeled lines or 2–4 sentences (https://picsart.com/blog/kling-ai-prompts/)
• Iterate with lock 4 slots, vary 1

FAQ

How do I choose the best AI video prompt formula for ads?

Use a formula that forces action and filming decisions: Subject → Movement → Scene → Camera → Lighting. Kling publishes a closely related structure in its quickstart guide. (https://kling.ai/quickstart/text-to-video-prompt-guide)

What’s the fastest way to stop “pretty but wrong” outputs?

Rewrite Movement so it’s straightforward and short-clip-friendly (one clear action). Kling explicitly recommends straightforward movement suitable for short videos. (https://kling.ai/quickstart/text-to-video-prompt-guide)

How should I describe camera movement without overcomplicating it?

Use natural camera language like close-up, wide shot, low angle, slow push-in, pan, tilt, tracking—and pick just one move. (https://kling.ai/blog/kling-ai-prompt-guide)

Should I write prompts as lists or as directions?

Directions tend to work better than lists for Kling 3.0 prompting guidance; treat your prompt like scene direction. (https://blog.fal.ai/kling-3-0-prompting-guide/)

How long should my prompt be for stable results?

A practical rule is 2–4 sentences (or labeled lines). This is cited as producing more stable, detailed outputs than one-line prompts. (https://picsart.com/blog/kling-ai-prompts/)

How do I scale this across lots of variations?

Keep the same 5-slot card, then produce variants by swapping one slot at a time (Camera, then Lighting, then Scene). For programmatic generation, Veo3Gen offers a developer API.

Generate variations faster with Veo3Gen

If your bottleneck is producing consistent variants (different hooks, angles, and 9:16 vs 16:9), Veo3Gen is an affordable way to access Google’s Veo 3.1 video models without Google’s enterprise pricing. You can start with free credits, iterate cheaply in Lite, default to Fast, and export finals in Quality—with native synchronized audio generated in one pass.

Closing CTA: When you’re ready to systematize your ad workflow, use Veo3Gen’s modes for the right stage (preview → iterate → finalize) and reuse the same 5-slot card to keep your creative consistent across every output.

Start creating with Veo3Gen

Veo3Gen gives you affordable Veo 3.1 video generation with native audio, up to 4K, and credits that never expire — with free credits to start.

• Generate your first video now: Get started
• Compare plans and pay-as-you-go pricing: See pricing