How to write a Veo 3 prompt
Veo 3.1 reads a prompt like a shot list. It does its best work when every sentence tells it one thing: who is in the frame, what they do, where they are, how the camera sees it and what the light looks like. The generator above writes the prompt in that order so nothing important is left for the model to guess.
- Subject and action first. “A baby hippo in ski goggles skis downhill” beats “ski video with a hippo”. Name one main subject and one clear action.
- Then the place. A specific setting (“a sunny alpine slope with fresh powder”) gives Veo the colors and the scale.
- Then the camera. Shot size plus one movement. Stacking three moves in eight seconds usually gives a mushy result.
- Then the look. Lighting and style words decide whether the clip feels like a film, a phone video or security footage.
- Sound last. Put spoken lines in quotes and list the sounds you want to hear. Veo treats quoted text as dialogue.
Example Veo 3 prompts
Casual vertical phone footage, like a real social media post. A first-person POV shot of a baby hippo in ski goggles skiing downhill, then wiping out into deep powder, on a sunny alpine ski slope. Handheld camera with natural shake. Lit by warm golden hour sunlight. Audio: wind rushing past, skis carving snow. A voice behind the camera says: "Did anyone else see that?" No subtitles.
Cinematic film look, shallow depth of field, subtle film grain. A close-up shot of a hand lifting the lid off a matte black box to reveal a glowing watch, on a dark marble table. The camera slowly pushes in. Lit by clean soft studio lighting. Audio: a soft click, then a low synth swell. No subtitles.
Want ideas that already work on TikTok? The trending AI prompt library has tested prompts for CCTV, GoPro, rescue and action figure trends.
Veo 3.1 settings in ShortStudio
In the ShortStudio app, Veo 3.1 makes 4, 6 or 8 seconds clips in 9:16 or 16:9 at a flat price, whatever the length. ShortStudio runs the fast version of Veo 3.1 without generated sound, so keep the audio lines if you plan to use the prompt in a tool that creates Veo audio, or switch to Kling 3.0 for native sound inside the app.
| Quality | Price | One video |
|---|---|---|
| Default | 38 credits per video | 38 credits |
See how Veo compares with every other model on price in the AI video cost calculator.
When to pick another model
- Longer clips or 4K: Kling 3.0 goes up to 15 seconds and has a 4K option.
- Several shots in one clip: Seedance handles cuts between shots well.
- A still photo first: start with Nano Banana, then animate the photo.