Veo 3 Prompt Builder

Start from a preset

1 · Subject

Who or what is in the video? Leave blank and we’ll use a default creator.

2 · Action

3 · Setting

4 · Camera movement

5 · Lighting

6 · Duration feel

Veo clips are ~8s — this steers the pacing and camera speed inside the clip.

Dialogue (optional)

Keep it under 10 words — short lines lip-sync best. Hindi works great.

01What this builder does

Veo 3 is the most literal-minded video model Google has shipped: it follows structured prompts beautifully and improvises badly when you’re vague. “A guy in a café” gets you a random guy in a random café. A prompt that names the subject, the action, the setting, the camera move and the light — in that order — gets you the shot you actually pictured. This builder assembles exactly that structure for you: a clean subject-action-setting sentence, then camera and lighting direction, a pacing line for how the clip should feel, an optional spoken dialogue line formatted the way Veo likes it, and a quality line that suppresses the usual Veo artefacts (garbled text, watermarks, wobbly faces). The three presets load complete proven setups — including the Talking Food format behind half the food Reels on Indian Instagram right now — so you can ship a known-working scene first and customise from there.

02Tips to get the best result

One action per clip. “Walks in, sits, orders, laughs” confuses Veo — give it one clear verb and let the camera do the storytelling.
Keep dialogue under ten words and say the language in the prompt (the builder does this for you). Long lines drift out of lip-sync.
Never ask Veo to render text — signboards, captions and logos come out garbled. Add text later in CapCut or your editor.
Shoot vertical ideas vertically: choose 9:16 in Flow/Gemini before generating. Cropping a 16:9 Veo clip for Reels wastes half the frame.
If faces wobble, simplify the camera. A static shot with a strong subject beats a drone move with a melting face, every time.

03How to use it (3 steps)

Tap a preset to load a proven scene, or fill in the fields yourself — subject, action, setting, camera, lighting, pace.
Tap Build Veo 3 prompt and copy the assembled prompt.
Paste it into Google Flow or the Gemini app (Veo), pick your aspect ratio, and generate. Want a photo prompt instead? Use the AI prompt generator.

04Frequently asked questions

Where do I paste the Veo 3 prompt this builder creates?−

In Google Flow (labs.google/flow) or the Gemini app with Veo selected. Paste the full assembled prompt as-is — the structure (subject, action, setting, camera, lighting, negative line) is written in the order Veo 3 responds to best. Choose your aspect ratio (16:9 or 9:16) in the app before generating.

How long are Veo 3 videos, and what does ‘duration feel’ change?+

Veo 3 generates clips of about 8 seconds in the Gemini app. The duration-feel setting doesn’t change the file length — it changes the pacing language in your prompt (‘fast, snappy beat’ vs ‘slow, lingering take’), which steers how Veo times movement and camera speed inside the clip. For a 15-second story, generate two clips and stitch them in any editor.

How do I make my Veo 3 character speak Hindi dialogue?+

Use the dialogue box and keep the line under ten words — short lines lip-sync far more accurately. The builder formats it as a spoken line with the language specified. Avoid asking for on-screen text or subtitles in the same clip; Veo still garbles rendered text often, so add captions later in your editor.

What do the three presets do?+

They load a complete, proven scene setup into the builder in one tap: the viral Talking Food format (a dish complaining to camera in Hindi), a Monsoon Romance couple shot, and a clean Product Spin ad. Use them as-is, or tweak one field — a different setting or camera move — to make the format yours.

Updated on 2 October 2026 · Free to use