Veo 3.1 “Meta Prompts” for Creators: A Copy‑Paste Director Prompt That Writes Your Shot Prompts (Plus 5 Guardrails to Stop Slop) (as of 2026-05-30)

What a “meta prompt” is (and when it beats hand‑writing prompts)

A meta prompt is a “prompt that writes prompts.” Instead of writing every shot prompt from scratch, you write one director-level instruction that:

• collects a small brief (offer, audience, brand rules, assets)
• outputs a consistent shot list + per-shot prompts
• generates platform variants (9:16, 1:1, 16:9)

This is especially useful with Veo 3.1 because Google frames prompting as a move from simple generation to creative control—a directing mindset rather than a one-line request. (https://cloud.google.com/blog/products/ai-machine-learning/ultimate-prompting-guide-for-veo-3-1)

Use a meta prompt when you’re:

• producing weekly content for the same brand
• making multiple deliverables from one offer (UGC + demo + teaser)
• trying to keep camera language, style, and tone coherent across outputs

The Director Meta Prompt (copy/paste)

Google Cloud’s Veo 3.1 prompting guidance emphasizes directing with clear intent and creative controls. Veo 3.1 is also described as a state-of-the-art video generation model with professional-grade creative controls, multiple aspect ratios, and rich synchronous audio. (https://cloud.google.com/blog/products/ai-machine-learning/ultimate-prompting-guide-for-veo-3-1)

Paste this as your “director meta prompt” in your workflow. It’s written so you can reuse it every week.

Director Meta Prompt (Veo 3.1)

Copy/paste:

You are “Director Promptwright” for Veo 3.1.

Goal:
Generate on-brand, production-ready Veo 3.1 shot prompts from the brief below.

Brief (fill in):
1) Offer: [WHAT IS BEING SOLD + 1 key benefit + 1 proof point]
2) Audience: [WHO + context + pain/desire]
3) Platform(s): [TikTok/IG Reels/YouTube Shorts/YouTube/etc.]
4) Deliverables needed: [UGC ad / product demo / social teaser / etc.]
5) Brand voice: [3–5 adjectives + banned tones]
6) Visual style / lookbook:
   - References: [describe 2–3 visual references in words; no links required]
   - Color palette: [list]
   - Wardrobe/props: [list]
   - DO NOT DO: [list]
7) Assets provided:
   - Product images: [yes/no + description]
   - Logos/packaging: [yes/no + usage rules]
   - First frame / last frame constraints (optional): [describe]
8) Must-include messages: [bullets]
9) Must-avoid topics/claims: [bullets]
10) Accessibility: [captions on/off, readable super rules]

Global constraints (non-negotiable):
- Output three separate variants: 9:16, 1:1, 16:9.
- Produce a numbered shot list with 6–10 shots.
- For each shot, provide:
  (a) Shot name + duration estimate
  (b) Camera language (framing, lens feel, movement)
  (c) Subject/action + environment
  (d) Lighting + color
  (e) Audio fields: SFX, music vibe, dialogue/VO (if any), on-screen text
  (f) A single “Veo prompt” paragraph that a creator can paste as the shot prompt
  (g) Negative constraints / safety notes (only when relevant)

Prompt quality rules:
- Be specific. Ban vague words: “cinematic,” “high quality,” “beautiful,” “stunning,” “epic,” “viral,” “dynamic,” unless you define what they mean in concrete visual terms.
- Keep each shot’s Veo prompt <= 80 words.
- If constraints conflict, prioritize in this order:
  1) Must-avoid topics/claims
  2) Brand voice + lookbook DO NOT DO
  3) Must-include messages
  4) Platform fit
  5) Everything else
- Avoid chaotic camera moves. Prefer one clear move per shot.

Output format:
A) Creative summary (3 bullets)
B) Shot list for 9:16
C) Shot list for 1:1
D) Shot list for 16:9
E) Consistency notes: what must stay identical across all variants
F) A short “things to double-check” list (max 6 bullets)

Now generate the outputs.

How to fill the inputs in 3 minutes

Offer (30 seconds)

Write it like: “[Product] helps [audience] get [benefit] because [proof].”

Example: “A magnesium gummy that helps busy runners sleep deeper because it combines magnesium glycinate + L-theanine.”

Audience (30 seconds)

Include the viewing context: “scrolling at night,” “in a gym parking lot,” “commuting,” etc.

Brand voice + lookbook (90 seconds)

This is where you stop “generic AI video.” Don’t say “cinematic.” Instead specify:

• composition: centered product hero, clean negative space, or handheld selfie
• lighting: soft window light, bright retail fluorescents, warm lamp light
• texture: glossy studio, real home bathroom, outdoor golden hour

Assets + frame constraints (30 seconds)

If you have a hero product image or packaging, note it. Google Cloud highlights narrative control features like “first frame, last frame” in the Veo 3.1 context via partner commentary—use that idea when you do have specific start/end requirements. (https://cloud.google.com/blog/products/ai-machine-learning/ultimate-prompting-guide-for-veo-3-1)

Output spec that actually matters (shot list + camera + audio)

Veo 3.1 is positioned with rich synchronous audio and creative controls, so your meta prompt should always force audio fields even if you end up leaving some blank. (https://cloud.google.com/blog/products/ai-machine-learning/ultimate-prompting-guide-for-veo-3-1)

The shot-level fields in the meta prompt are deliberate:

• Camera language keeps motion readable (and helps you avoid random zooms).
• Lighting + color anchors brand consistency.
• Audio (SFX/music/VO/text) prevents “silent montage syndrome” and keeps your intent explicit.
• Negative constraints/safety notes reduce accidental claims or off-brand elements.

Worked examples: 1 offer → 3 deliverables

Below are tiny creator briefs and the exact meta‑prompt-filled version you can paste.

Example 1: One offer, UGC ad deliverable

Tiny brief

• Offer: Toothpaste tablets; benefit: travel-friendly; proof: TSA-friendly tin + no liquid.
• Audience: frequent flyers who hate leaky toiletry bags.
• Platform: IG Reels + TikTok.
• Brand voice: friendly, practical, not preachy.
• Lookbook: bright bathroom, handheld selfie, clean white + mint accents.
• Must-avoid: medical claims.

Meta prompt filled

1) Offer: Toothpaste tablets in a small tin; helps frequent travelers keep oral care simple with no liquid mess; proof: TSA-friendly tin and tablets (no liquid tube).
2) Audience: frequent flyers scrolling while packing; they want lighter bags and fewer leaks.
3) Platform(s): TikTok, IG Reels
4) Deliverables needed: UGC ad
5) Brand voice: practical, upbeat, concise, no-shame; banned tones: alarmist, medical, luxury-snob
6) Visual style / lookbook:
   - References: handheld selfie in a real bathroom; quick packing counter shots; close-ups of tablets dissolving
   - Color palette: white, mint, light gray
   - Wardrobe/props: casual tee, carry-on, toiletry bag, tin
   - DO NOT DO: dentist office vibes; clinical charts; exaggerated “before/after”
7) Assets provided:
   - Product images: yes, tin + tablets close-up
   - Logos/packaging: yes; keep logo facing camera in hero shots
   - First frame / last frame constraints (optional): first frame is bathroom mirror selfie; last frame is tin going into carry-on pocket
8) Must-include messages: TSA-friendly; no leaky tube; fresh mint taste
9) Must-avoid topics/claims: no cavity prevention claims; no “whitening in X days”
10) Accessibility: captions on; keep super text <= 6 words per beat

Example 2: Product demo deliverable

Tiny brief

• Offer: clip-on phone light for creators.
• Audience: small creators filming at night.
• Platform: YouTube (16:9) + Shorts.
• Lookbook: split-screen “before/after” style without making performance claims.

Meta prompt filled

1) Offer: Clip-on phone light; helps small creators get more even face lighting at night; proof: compact clip + adjustable brightness.
2) Audience: creators filming at home after work; they hate grainy, shadowy footage.
3) Platform(s): YouTube, YouTube Shorts
4) Deliverables needed: product demo
5) Brand voice: helpful, nerdy-simple, confident; banned tones: hypey, guru, mean
6) Visual style / lookbook:
   - References: desk setup at night; clean demo table shots; simple split-screen comparisons
   - Color palette: neutral blacks + warm skin tones
   - Wardrobe/props: hoodie, phone tripod, desk lamp
   - DO NOT DO: fake “miracle” language; unrealistic glow effects
7) Assets provided:
   - Product images: yes
   - Logos/packaging: no
   - First frame / last frame constraints (optional): none
8) Must-include messages: adjustable brightness; clips onto phone; portable
9) Must-avoid topics/claims: no quantified performance claims; no “studio quality guaranteed”
10) Accessibility: captions on; avoid tiny UI text

Example 3: Social teaser deliverable

Tiny brief

• Offer: indie game launch.
• Audience: cozy-game fans.
• Platform: TikTok.
• Lookbook: soft pastel, calm pacing, gentle camera moves.

Meta prompt filled

1) Offer: Cozy indie game launch; helps cozy-game fans unwind with short quests; proof: pastel world + gentle soundtrack vibe.
2) Audience: cozy-game fans scrolling at bedtime; they want calm, not adrenaline.
3) Platform(s): TikTok
4) Deliverables needed: social teaser
5) Brand voice: warm, whimsical, calm; banned tones: intense, sarcastic, chaotic
6) Visual style / lookbook:
   - References: pastel diorama look; slow pans; soft bloom lighting
   - Color palette: peach, lavender, cream
   - Wardrobe/props: none (game world only)
   - DO NOT DO: shaky cam; aggressive glitch cuts; horror shadows
7) Assets provided:
   - Product images: yes, key art
   - Logos/packaging: yes; logo only in final shot
   - First frame / last frame constraints (optional): first frame matches key art composition; last frame is logo on cream background
8) Must-include messages: wishlist now; cozy quests; release date card
9) Must-avoid topics/claims: no awards/press quotes
10) Accessibility: captions on; high-contrast text

5 guardrails to stop slop (and what to do when the model ignores them)

1) Conflict resolver (prevents impossible briefs)

Add (already included):

• “If constraints conflict, prioritize…”

If you still get contradictions, move the priority list above the brief so it’s seen earlier.

2) Length cap (prevents prompt bloat)

Concrete pattern:

• “Keep each shot’s Veo prompt <= 80 words.”

If outputs are still long, tighten it to 50 words and reduce shots from 10 to 6.

3) Banned vagueness list (prevents generic style)

Concrete pattern:

• “Ban vague words… unless you define what they mean in concrete visual terms.”

If the model keeps saying “cinematic,” add: “Replace ‘cinematic’ with lens + lighting + movement specifics.”

4) Style anchoring via lookbook + DO NOT DO

Concrete pattern:

• “References: [describe 2–3 visual references in words]”
• “DO NOT DO: [list]”

If you get off-brand imagery, expand DO NOT DO with 3–5 very visual bans (e.g., “neon cyberpunk,” “overcranked slow motion,” “lens flare everywhere”).

5) Camera sanity rule (prevents chaotic motion)

Concrete pattern:

• “Avoid chaotic camera moves. Prefer one clear move per shot.”

If shots still feel hyperactive, add: “Max camera move speed: slow; no snap zooms.”

Iteration loop: refine the meta prompt (not every shot)

When something’s off, don’t rewrite 8 shot prompts. Edit one line in the meta prompt and regenerate.

Examples of high-leverage tweaks:

• Too generic → add 2 more lookbook specifics + 3 more DO NOT DO bans.
• Wrong tone → add “banned tones” and a sample line of VO style (e.g., “short, matter-of-fact sentences”).
• Inconsistent across ratios → add a “Consistency notes” requirement (already included) and specify what must remain identical (hero prop, wardrobe color, tagline).

As of 2026-05-30, Google Cloud positions Veo 3.1 as stable and generally available for production on Vertex AI—so building a repeatable weekly prompting system is a reasonable workflow choice. (https://cloud.google.com/blog/products/ai-machine-learning/ultimate-prompting-guide-for-veo-3-1)

Mini checklist: before you generate, before you export

• One sentence offer + one proof point (no extra claims)
• 3–5 brand voice adjectives + 3 banned tones
• Lookbook: palette + lighting + environment + DO NOT DO list
• Must-include messages (bullets) and must-avoid claims (bullets)
• Decide deliverables and require 9:16, 1:1, 16:9 outputs

Troubleshooting (quick tweaks)

Outputs are too generic

Add to lookbook: “Use realistic environment details: [tile type, countertop material, time of day, wardrobe].”

Outputs are too long

Reduce: “6 shots only” and “<= 50 words per shot.”

Brand voice keeps drifting

Add: “Dialogue/VO must follow: [example sentence]. Avoid: [2 example sentences you hate].”

Camera moves feel chaotic

Add: “No more than 1 move per shot; default to locked-off tripod unless specified.”

FAQ

What makes this a “Veo 3.1 meta prompt” instead of a normal template?

It’s designed to generate shot prompts + platform variants from a single brief, aligned with Google’s emphasis on directing for creative control. (https://cloud.google.com/blog/products/ai-machine-learning/ultimate-prompting-guide-for-veo-3-1)

Do I need to generate audio directions too?

If you want more predictable results, yes—Google describes Veo 3.1 as having rich synchronous audio, so specifying VO/SFX/music intent is part of “directing,” not an afterthought. (https://cloud.google.com/blog/products/ai-machine-learning/ultimate-prompting-guide-for-veo-3-1)

Can this help me keep consistency across multiple videos?

It can improve consistency by standardizing your brief and shot fields. For longer continuity workflows, some guides discuss extending existing videos and starting from specific frames; treat those as optional techniques depending on your pipeline. (https://www.datacamp.com/es/tutorial/veo-3-1-complete-guide-with-examples)

If I use LTX Studio, do aspect ratios matter?

LTX Studio’s Veo 3.1 access mentions aspect ratio options like 16:9 and 9:16, so it’s practical to request platform-specific variants up front. (https://ltx.studio/blog/veo-prompt-guide)

Related reading

• /blog/veo-3-1-release-date-launch-notes
• /blog/what-is-veo-3-api
• /blog/veo-3-api-features-guide

Ready to generate shot prompts programmatically?

If you’re turning this into a repeatable content engine (brief in → shot list out → variants), it’s often easiest to wire it into your tooling.

• Explore the developer workflow on Veo3Gen API: /api
• Compare options and scale when you’re ready: /pricing