Veo 3 is available on Layer as a video generation model, currently offered to select enterprise customers. If you have access, you’ll be able to use Veo 3 directly within your Layer workspace as part of the video generation toolset.
What makes prompting for Veo 3 different: What sets Veo 3 apart from other video generation models is its ability to produce synchronized audio alongside visuals. This means you should consider not just what the viewer sees, but also what they hear. Your prompt can describe ambient noise, background music, or exact dialogue lines—each of which will be reflected in the final video. This enables a richer, more immersive viewing experience, especially when audio is used to reinforce emotion, context, or character speech.
Prompt Structure
A great Veo prompt includes:
Subject – What or who is in the scene.
Action – What they’re doing.
Environment – Where it’s happening.
Style – The visual tone or genre.
Camera – Angle, movement, and framing.
Mood – The emotional tone or lighting.
Audio – Sound effects, music, dialogue.
Examples:
These all include first-frame reference images, which are recommended when trying to create something with a specific style.
  | camera stays steady, the card pack rips open a the top and theres a flash of light and 6 trading cards come out and line up on the screen face down, the back of the cards has a colorful logo and says "Snek", there is the sound of the card pack ripping and then sound effect of the cards flying out |
  | idle animation for a video game lobby, shuffling slightly side to side, drop gun down to the side |
  | Running away from the green creatures as the camera backs up, anime style, the green monsters chasing, you can hear the characters running steps and panting |
  | camera pans up and then follows the character as he runs forward down a video game style jungle path, there are blocks he needs to jump over as he runs, sound of him running and jumping |
Tips:
Be precise with sound descriptors.
Use clear, quotable dialogue for lip-sync moments.
Pair emotional tone with musical or atmospheric cues.
Make sure your audio cues support the visual story.
Best Practices
Avoid Sensitive Language: Veo 2 and Veo 3 are sensitive to certain keywords that may trigger safety filters and block generation. If a video fails to generate, consider whether any terms in the prompt might be flagged. For example, replacing "boy" with "character" or "child" may help bypass unintended censorship while maintaining your creative intent.
Iterate Prompt-by-Prompt: Refine your video in stages. Start with the base scene. Add complexity (camera, lighting, effects, sound) in layers.
Maintain Consistency: For multi-shot sequences or recurring characters, copy and adapt prompt elements to maintain coherence.
Review and Regenerate: Evaluate results carefully. If a shot feels off, adjust the prompt—clarify actions, simplify language, or add detail.
Use Negative Prompts: Use negative prompts. The Layer app includes a dedicated negative prompt field in the video generation prompt box—use it to list any elements you want to exclude. For example: "no text," "no subtitles," or "no logos on screen." This helps maintain clean compositions and prevent unwanted distractions.
Match Sound to Scene: Specify ambient sounds or dialogue when needed. Veo supports native audio but benefits from detailed audio descriptions.