Prompts for Veo: Image-to-Video & Audio

01 — Overview

What Veo rewards

Google’s Veo family handles both text-to-video and image-to-video, and recent versions can generate audio alongside the picture. For commercial work the image-to-video path is usually the safer one: you approve the composition as a still, then ask Veo for motion only, which removes most identity and layout risk.

Because audio is generated too, treat it as part of the brief. Say what ambience you want, whether there is dialogue, and what must not appear — stray music beds and unwanted voices are common surprises otherwise.

01Image-to-video animation that respects an approved frame.
02Native audio generation, including ambience and dialogue.
03Good camera-move adherence when the move is described precisely.
04Available through Google’s developer surfaces alongside Gemini.

02 — Anatomy

What a Veo prompt should cover

Source Text-to-video or an attached image. Say which, and what to preserve.
Motion brief Primary motion, secondary motion, and what stays perfectly still.
Camera Move, direction, amount and speed. Include "no shake" if you want stability.
Light + grade Constant or changing, and how.
Audio Ambience, foley, dialogue, and explicit exclusions like "no music".
Duration Clip length, and whether it should loop.
Do-not-change list Faces, logos, text, product geometry, composition.

03 — The prompts

4 prompts written for Veo

Paste, then replace the [bracketed] parts. Keep the wording identical for anything you want repeated across a set.

Filter 4 prompts

Nothing matches that word. Clear the field to see every prompt.

Prompt 01

Image-to-video brand animation

Animating an approved still without redesigning it.

image to video
brand

Template · plain text

[Attach the approved still image]

Animate the attached image. Preserve the composition, colours, typography and every design element exactly. Do not redraw, restyle, reframe or crop.

MOTION
- Primary: [element] [small specific movement, e.g. steam rising slowly].
- Secondary: [element] [subtle movement, e.g. light flickering gently].
- Camera: extremely slow push in, under 4% scale change over the clip, no shake.

HOLD PERFECTLY STILL
- Logo, product geometry, any text, and all faces.
- Lighting direction, colour grade and contrast: unchanged.
- No new objects, no morphing, no style drift, no added lens flare.

DURATION: 5 seconds, loopable if possible.
AUDIO: [quiet room ambience] only. No music, no voices, no whooshes.

Prompt 02

Text-to-video shot with audio

Generating picture and sound in one pass.

text to video
audio

Template · plain text

Medium shot, 8 seconds, one continuous take.

SUBJECT: [subject], [wardrobe/material detail], [state].
ACTION: [one continuous action completing within 8 seconds].
CAMERA: handheld follow at walking pace, subject held in the centre third, gentle natural movement, no whip pans.
ENVIRONMENT: [location], [time of day], [weather], [two background details].
LIGHT: [source] from camera-right, soft, constant exposure across the clip.
LENS: 35mm, f/2.8, shallow but not extreme depth of field.
GRADE: naturalistic, mild contrast, fine grain.

AUDIO
- Ambience: [specific environment sound], quiet and continuous.
- Foley: [footsteps on gravel / fabric movement], matched to the action.
- Dialogue: none.
- Exclude: music, crowd noise, sirens, birdsong.

No cuts, no captions, no text overlays, no additional people.

Prompt 03

Explainer clip with spoken line

A short presenter or voice moment where sync matters.

dialogue
explainer

Template · plain text

Medium close-up, 7 seconds, static camera on a tripod.

SUBJECT: [age] [description] presenter, [wardrobe], relaxed posture, natural blinking and micro-movements.
DIALOGUE (spoken exactly, conversational pace, [accent]): "[one or two sentences, max 22 words]"
DELIVERY: [calm and clear], eye contact with the lens, one small open-hand gesture on the final phrase.
FRAMING: eyes on the upper third, modest headroom, subject centred.
ENVIRONMENT: [setting] softly out of focus, [one background object].
LIGHT: soft key camera-right, gentle fill, background one stop darker.
LENS: 50mm, f/2.5.

AUDIO: clean voice, quiet [room tone], no music, no reverb.

No captions, no on-screen text, no cuts. Keep lip movement matched to the dialogue and do not extend the line.

Prompt 04

Loopable background plate

Website headers and app backgrounds that play continuously.

loop
background

Template · plain text

Abstract background plate, 6 seconds, seamless loop with matching first and last frames.

CONTENT: [soft flowing gradient / drifting particles / slow-moving fabric] in [#hex] and [#hex] only.
MOTION: continuous, even speed, no beginning or end emphasis, no direction changes.
CAMERA: static.
DEPTH: gentle blur variation for depth, no sharp focal point that draws the eye.
BRIGHTNESS: constant, mid-tone, safe for white text overlaid on top — keep contrast low.
GRADE: clean, no grain, no vignette.

AUDIO: none.

No recognisable objects, no faces, no text, no logos, no strobing or high-frequency flicker.

04 — Craft notes

What separates a good prompt from a wasted run

Do this 05

Prefer image-to-video for anything brand-critical.
List explicitly what must not change.
Direct the audio, including what to exclude.
Quantify camera movement ("under 4% scale change").
Keep dialogue short so sync holds.

Not this 05

Asking for a restyle and a motion pass at once.
Leaving audio unspecified and being surprised by music.
Overlaying critical text in the generation instead of in post.
High-contrast background plates behind body copy.
Long monologues in a single clip.

05 — Questions

Asked often, answered plainly

Text-to-video or image-to-video?

Image-to-video whenever composition, product accuracy or identity matters, because you lock the frame before motion enters the picture. Text-to-video is better for exploration and atmosphere where the exact frame is negotiable.

How do I stop unwanted music appearing?

Name the audio you want and list exclusions explicitly: no music, no voices, no sound effects beyond the foley you asked for. Silence is also a valid instruction.

Can Veo render my logo correctly?

Do not rely on it. Generate the motion without the logo and composite the real asset in post, or start from a still that already contains it and freeze that region with a do-not-change instruction.

06 — Sources

Read the primary documentation

Vendors change flags, limits and defaults often. Where this site disagrees with official docs, the docs win.

Prompt for Veo

What Veo rewards

What a Veo prompt should cover

4 prompts written for Veo

Image-to-video brand animation

Text-to-video shot with audio

Explainer clip with spoken line

Loopable background plate

What separates a good prompt from a wasted run

Do this 05

Not this 05

Asked often, answered plainly

Read the primary documentation

Browse by output instead

Video prompts

Audio prompts