A free visual dictionary of camera-movement and motion keywords — every word paired with a real generated video, so you see exactly how it moves before you prompt. Model-agnostic vocabulary for Sora, Veo, Kling, Runway, Seedance, Hailuo, Midjourney & more.
AtomWordsThe Visual Dictionary of AI Art Keywords
AI video prompt keywords are single, atomic terms — like "dolly zoom", "whip pan", "orbit shot", or "slow motion" — that each control one aspect of an AI-generated video: camera movement, motion, pacing, or cinematography. Combining a few precise keywords gives far more controllable results than one long paragraph prompt.
A camera movement prompt names how the virtual camera moves through a shot — dolly in/out, truck, pedestal, pan, tilt, crane, orbit, or handheld. Naming the exact movement (e.g. "slow dolly in") is far more reliable than vague direction like "cinematic", and each keyword here links to a real generated clip so you can see the motion before you prompt.
The keywords are model-agnostic cinematography vocabulary, so they carry across the major AI video generators — Sora, Veo, Kling, Runway, Seedance, Hailuo, Midjourney, Pika, and Luma. Each keyword's page shows a real generated example so you can judge how the effect reads before adapting it to your model.
Yes — browsing is free, and every keyword entry published so far is fully free: watch the example clip, copy the complete prompt, and download the media, including for commercial use. See the License page for the exact terms.