A prompt checklist for text-to-video: camera, light, action, place#

Most weak AI clips come from prompts that only name a subject. "A dog on a beach" leaves the model to guess everything else. Before you hit generate, run through four questions.

1. Camera. Where is the viewer? Low angle, eye level, a slow push in, a wide establishing view?
2. Light. Golden hour, overcast, neon signs, a single desk lamp? Light sets mood faster than adjectives.
3. Action. What changes during the shot? A head turn, waves rolling in, a door swinging open.
4. Environment. Weather, background detail, time of day.

A rewritten prompt: "Low-angle shot of a wet terrier shaking off on a windy beach at golden hour, gulls in the background, waves breaking left to right."

In the AI video generator on videofx.io, text prompts render at 1080p, and clips of 5-10 seconds usually arrive in about 30-45 seconds with ambient sound already matched to the scene. New accounts get 10 free credits, so you can test the checklist on a couple of drafts.

VideoFX is an independent AI video studio, not affiliated with Google, TikTok or other products named VideoFX.