A tool that builds AI text-to-video prompts the way a film shot is described. Type your subject / scene (e.g. an astronaut on a red desert) and the action / motion (e.g. walking toward the horizon), then just pick from dropdowns: shot (wide, close-up, low angle…), camera move (dolly in, tracking, aerial drone…), lighting (golden hour, neon, backlit…), style (cinematic, anime, vintage 8mm…), mood (epic, dreamy, suspenseful…), speed (slow motion, time-lapse…), aspect ratio (16:9, 9:16…) and length (5 s, 10 s…). It joins them into one natural-English sentence — "[style] [shot] of [subject], [action], [camera], [lighting], [mood], [speed]. [aspect], [length]." — in the form models like Sora, Kling, Runway, Luma Dream Machine and Pika read best. These models respond strongly to English cinematography terms (dolly in, golden hour, 16:9), so the output is in English. The subject, action and extra-detail fields are free text and only the options you select are added (unselected ones are skipped), so you can keep a prompt minimal or pile on detail. Your camera / lighting / style choices are saved in the browser so you start from the same defaults next time (the text you type is not stored). All assembly is plain string work in your browser — the ideas you type and the prompt it produces are never sent to any API or server.