Google’s own guidance for Veo boils down to seven elements, and the examples on this page use all of them: subject, action, context, style, camera position and motion, ambiance, and audio. Write it like a shot brief: “Cinematic aerial drone shot (camera) of turquoise waves breaking against limestone cliffs (subject, context), slowly pushing forward along the coastline (motion), golden hour haze (ambiance)”. Camera vocabulary is the biggest lever — dolly in, tracking shot, aerial, low-angle, close-up, rack focus — because one camera phrase changes the entire clip.
Audio is prompted in the same text: describe ambient sound and effects in plain words, and put dialogue in quotes — Veo 3.1 lip-syncs short lines convincingly. On Veo, audio is optional and priced separately; on Gemini Omni Flash sound is always on and the model also picks the clip length itself. Silent Veo 3.1 clips start at $0.10 on the Lite tier; every example below is labeled with the model and what that exact generation cost.