On this page
- How do you write a Veo 3 prompt?
- How do you write dialogue and sound in Veo?
- Veo 3 prompt for a UGC selfie ad
- Veo 3 prompt for a podcast clip ad
- Veo 3 prompt for a street interview ad
- Veo 3 prompt for a short drama ad
- Veo 3 prompt for a claymation or animation ad
- Veo 3 prompt for a product shot ad
- Veo 3 prompt for a brand film ad
- How do you keep a person or product consistent in Veo?
How do you write a Veo 3 prompt?
Write a Veo prompt in five parts: camera, subject, action, context, then style and ambiance. That is the formula in Google Cloud's prompting guide for Veo 3.1, and it starts with the shot, not the person.
The Gemini API Veo guide lists the same elements in a different order: subject, action and style as core, then camera positioning and motion, composition, focus and lens effects, and ambiance as optional. Both guides ask for descriptive sentences rather than lists of tags.
| Element | What to write for an ad | Example terms from Google's guides |
|---|---|---|
| Camera | Shot size, angle, move | Dolly shot, eye-level, close-up, two-shot |
| Subject | Who or what, with 2 to 3 visible details | A person, product or scene |
| Action | One thing they do | Turning their head, walking |
| Context | Location and time | The environment and background |
| Style and ambiance | Look, light, colour | Warm tones, film noir, cartoon |
Searches say "Veo 3", but the current model is Veo 3.1; the Gemini API marks Veo 3 as deprecated. The rules here are from the 3.1 guides.
How do you write dialogue and sound in Veo?
Put speech in quotation marks, and describe every sound in words. The Gemini API guide says to use quotes for specific speech, to describe sound effects explicitly, and to describe the ambient soundscape.
Google Cloud's guide adds labels: SFX: before a sound effect and Ambient noise: before the background sound. For negatives, it says to describe what you want, such as a street with no cars, rather than writing "no cars".
Duration (4, 6 or 8 seconds) and aspect ratio (16:9 or 9:16) are API parameters, not prompt text. Clips must be 8 seconds with reference images, 1080p or 4k, per the Gemini API docs. Plan one line of dialogue per 8-second clip.
The templates below are written for this page and have not been run through Veo; no outputs are shown. Square-bracket items are placeholders. For specs and joining clips, see Veo 3 for ads. To fill a template from a form, use the AI ad prompt builder.
Veo 3 prompt for a UGC selfie ad
A Veo UGC selfie prompt opens with the phone framing, then the person, the action and one quoted line. See the get-ready UGC examples.
Handheld front-camera selfie, medium close-up, slight natural shake.
A [age range] man with [hair] in a [hoodie], using the reference image
of [product]. He holds the product next to his face and taps it with one finger
in a [gym changing room / car interior] in the morning.
He says, "[hook line, under 15 words]"
Natural phone-camera look, flat daylight.
SFX: a locker door closing somewhere behind him.
Ambient noise: distant voices and a fan humming.
Veo 3 prompt for a podcast clip ad
Generate one host per clip and cut between them; Google does not publish a speaker limit, and one speaker per 8 seconds is easier to control. See the AI podcast ad examples.
Medium close-up, eye-level, static camera, shallow focus.
A woman in her [age] with [hair], over-ear headphones on, at a broadcast
microphone. She turns from her co-host toward the mic and lifts [product]
into frame, in a podcast studio with a lit bookshelf behind her.
She says, "[the line that names the product and the result]"
Warm tones, soft key light from the left.
Ambient noise: quiet room tone, a chair creaking once.
For the second host, keep the camera line and lighting identical and change only the subject and the quote.
Veo 3 prompt for a street interview ad
Street interviews use Google's speaker-label pattern so the interviewer and the passer-by stay separate. See the AI street interview examples.
Handheld medium shot, eye-level, documentary style.
A [age range] woman with [shopping bag], stopped on a busy [city] pavement.
A microphone with a foam cover enters from the left edge of frame.
Interviewer (off camera): "[quick question]"
Woman: (laughing, pointing at [product] in her bag) "[short answer]"
Bright overcast daylight, natural colour.
Ambient noise: traffic, footsteps, a bus pulling away.
Veo 3 prompt for a short drama ad
A short drama clip in Veo is one shot, one exchange. Write the delivery in parentheses before each quote, as Google's dialogue example does. See the AI drama ad examples.
Two-shot, static camera, slightly low angle.
A tired father in a [work shirt] and his teenage daughter at a kitchen table
at night. He pushes [problem object] away; she slides [product] across to him.
Daughter: (dry, half-smiling) "[line that names the fix]"
Father: (surprised, picking it up) "[short reply]"
Cinematic, warm tungsten light, soft shadows.
SFX: the product sliding on wood.
Ambient noise: a fridge humming.
Veo 3 prompt for a claymation or animation ad
Name the animation style first; the Gemini API guide treats style as a core element. See the claymation and craft examples or the Pixar-style examples.
Close-up, slow dolly in.
Claymation style. A small clay [animal] character with button eyes
sits inside a clay [lunchbox / backpack]. It lifts the lid of [product],
sniffs it, and its ears pop up, on a clay [school desk] scene.
The character says, "[one-line benefit]"
Handmade textures, visible fingerprints, soft studio light, [palette].
SFX: a soft pop as the lid opens.
Veo 3 prompt for a product shot ad
Product shots work well with Veo's first and last frame control: set the opening and closing images, and describe only the move between them. See the AI product demo examples.
Macro close-up, slow 90-degree arc around the product.
[product] from the reference image standing on [wet slate / brushed steel],
a thin layer of frost melting off its surface as the camera moves,
in a dark studio set.
Commercial look, hard side light, deep shadows, sharp focus on the label.
SFX: a faint crackle of ice.
Ambient noise: near silence.
Veo 3 prompt for a brand film ad
Brand-film prompts carry no dialogue; sound and light do the work. Google's guide also documents timestamp prompting for several shots in one generation. See the AI brand film examples.
Wide crane shot rising slowly.
A lone swimmer in a [colour] cap walking into a cold [lake / sea] at dawn,
holding [product], with pine trees and low mist behind.
Cinematic, cool blue tones, soft first light, slight film grain.
SFX: water lapping at her ankles.
Ambient noise: birds waking, a light wind.
How do you keep a person or product consistent in Veo?
Use reference images. The Gemini API docs say Veo 3.1 takes up to three asset images of one person, character or product and keeps that subject's appearance, and Google Cloud recommends making those references with Nano Banana. See Nano Banana prompts for ads for still-image templates.
For one script in every format, read one script, every format. If your Veo ads already perform on paid traffic, see how Anglr works for creators.
Who this is for
- Independent AI creators and editors. You make the ads yourself. You know your work performs, so you'd rather be paid on what it does than on a flat fee.
- AI agencies and studios. You run a team of editors. You write the brief, your team makes the versions, and you're paid on performance because you know your team's ads scale.
Not for you if
- You only want a flat fee.
- You're new to editing and not yet confident your ads will perform.
- You want every shot spelled out for you.
- You make one style and won't try others.
FAQ
How do you write dialogue in a Veo 3 prompt?
Put the exact words in quotation marks after who says them, for example: She says, "This one lasts all day." Google's Gemini API guide gives this rule, and its examples also use a speaker label before the quote.
How do you add sound effects to a Veo prompt?
Describe the sound explicitly in its own sentence. Google Cloud's Veo 3.1 guide labels these with SFX: and the background soundscape with Ambient noise:.
What is the Veo 3 prompt formula?
Google Cloud's guide gives Cinematography + Subject + Action + Context + Style and ambiance. The Gemini API guide lists subject, action and style as core, with camera, composition, focus and ambiance as optional.
Can I write negative prompts for Veo?
Google's guide recommends describing what you want instead, for example a quiet street with no people rather than no people. State the scene positively.
How many reference images can Veo 3.1 use?
Up to three images of a single person, character or product, per the Gemini API docs. Veo keeps that subject's appearance, and the clip must be 8 seconds when you use them.
Is Veo 3 the same as Veo 3.1?
Searches still say Veo 3, but on the Gemini API Veo 3 is marked deprecated and the current models are Veo 3.1, 3.1 Fast and 3.1 Lite. The prompts on this page follow the Veo 3.1 guides.
Sources
- Google Cloud – The ultimate prompting guide for Veo 3.1
- Google AI for Developers – Generate videos with Veo (Gemini API)
Checked: 12 Oct 2026