VODOOVeo 3 prompts: how to write for people, speech and sound
By VODOO AI · · 4 min read
Veo's strength is people: a person talking to camera with the right lip movement, the room's sound around them, and the small movements that make someone look alive. Most Veo prompts waste that by describing a scene and forgetting the sound and the speech.
Here's how to write for Veo: the order it reads best, how to write dialogue and sound, and prompts to copy for ads, UGC and short scenes.
The structure Veo reads best
- Who: a casting-note description of the person.
- Where and when: the place and the light.
- What they do: one action, then what they say, in quotes.
- How it sounds: the room, the background, the voice's tone.
- How it's filmed: one camera move, the lens, the look.
A man in his late twenties with a short beard and a grey hoodie sits on a kitchen stool in a small flat at night, warm lamp light. He leans towards the camera and says, quietly excited: "Okay, I didn't expect this to actually work." Fridge hum and a distant siren outside. Handheld phone framing, slight movement, natural skin.Make this in VODOO →
Writing dialogue
Put the exact words in quotes and say how they're said: tired, amused, whispering, mid-laugh. Keep it to one or two short sentences per clip; long speeches rush or drift. If two people talk, name who says which line.
Two friends in their twenties sit on a park bench in autumn, one holding two coffees. She hands one over and says, teasing: "You owe me for this." He takes it and replies, deadpan: "Put it on my tab." Leaves rustling, distant children playing. Static wide shot, then a slow push-in.Make this in VODOO →
Sound is half the prompt
Veo makes the sound with the picture, so describe it. Name two or three sounds that belong to the place. Silence is a choice too: "no music, only room tone" stops it from adding a soundtrack you didn't want.
Café
- Sounds to name
- Milk steamer, cups on saucers, low chatter
Street at night
- Sounds to name
- Distant traffic, footsteps, a passing car
Bathroom
- Sounds to name
- Tap running, small echo, a drip
Studio
- Sounds to name
- Room tone only, no music
| Place | Sounds to name |
|---|---|
| Café | Milk steamer, cups on saucers, low chatter |
| Street at night | Distant traffic, footsteps, a passing car |
| Bathroom | Tap running, small echo, a drip |
| Studio | Room tone only, no music |
UGC and ads with Veo
For UGC-style ads, ask for a phone camera: handheld, slightly imperfect framing, natural light, and a person talking like they would to a friend. Say what the product is and where it sits in the frame.
Vertical 9:16 selfie video: a woman in her forties with reading glasses on her head holds a small amber serum bottle labelled 'LUMEN' up to the phone in her bathroom, morning light, and says honestly: "Three weeks in, and my skin actually feels calmer." Tap running quietly behind. Handheld, natural skin texture, no beauty filter.Make this in VODOO →
More on this in AI UGC ads and the AI ad generator guide.
Common mistakes, and the fix
The lips don't match the words
- Why
- The line is too long
- Fix
- One or two short sentences per clip
A soundtrack appears
- Why
- Music wasn't ruled out
- Fix
- Write 'no music, only room tone'
The person looks airbrushed
- Why
- No texture was named
- Fix
- Name pores, stubble, flyaway hair, real light
The voice sounds flat
- Why
- No tone was given
- Fix
- Say how the line is said: tired, amused, quiet
| What happens | Why | Fix |
|---|---|---|
| The lips don't match the words | The line is too long | One or two short sentences per clip |
| A soundtrack appears | Music wasn't ruled out | Write 'no music, only room tone' |
| The person looks airbrushed | No texture was named | Name pores, stubble, flyaway hair, real light |
| The voice sounds flat | No tone was given | Say how the line is said: tired, amused, quiet |
Veo also responds well to a small action before the line: a breath, a glance away, picking something up. It gives the speech a reason to start and makes the person feel caught rather than posed.
A chef in a white jacket in a busy restaurant kitchen wipes her hands on a towel, glances at the pass, then looks to camera and says, slightly breathless: "Friday nights. Every single week." Pans sizzling, orders being called. Handheld, warm overhead light, steam in the air.Make this in VODOO →
Using Veo in VODOO
Open Veo in VODOO, or write your brief in Create and let the studio choose Veo when a shot needs a person speaking. To compare it with others on the same brief, see Veo vs Sora vs Kling. Start with 300 free credits a month.
Questions people ask
- How do I write a good Veo 3 prompt?
- Describe the person, the place and light, one action, the exact words in quotes with how they're said, the sounds of the room, and one camera move.
- Can Veo 3 make people talk?
- Yes. Put the dialogue in quotes and say the tone. Keep each clip to one or two short lines for the most natural lip movement.
- How do I stop Veo adding music?
- Say so in the prompt: 'no music, only room tone', and name the background sounds you do want.