Direct AI video generation like a cinematographer — camera, motion, and scene control.
Runway Gen-3 is a professional-grade video generation tool that responds to cinematographic language. Unlike image generation, video prompts must describe motion, camera movement, and temporal progression — static descriptions produce static-looking videos. The most effective Runway prompts read like a film director's shot description: camera movement, subject action, and atmospheric conditions all in one sentence.
Each prompt is annotated with the reasoning behind its structure.
Slow dolly forward through a rain-soaked Tokyo alley at night, neon signs reflecting in puddles, steam rising from a street food stall, a lone figure in a yellow raincoat walking away from camera, cinematic 24fps, anamorphic lens flare, shallow depth of field, muted colors with neon highlights
Camera movement (slow dolly forward) is specified first — Runway prioritizes the first instruction. Subject action (figure walking away) and environmental motion (steam, rain) give the scene multiple motion layers. Cinematographic specifics (24fps, anamorphic) set the visual register.
Time-lapse: storm clouds building over a mountain range, camera locked off on tripod, clouds rolling and churning from left to right, lightning flashes in the distance, golden hour light transitioning to dramatic purple dusk, 4K, photorealistic
'Camera locked off on tripod' is crucial — it tells Runway the movement should come from the clouds, not the camera. Multiple motion elements (clouds rolling, lightning, light changing) give the model clear direction for what should animate. Time-lapse framing sets pacing expectations.
Smooth 360-degree orbit around a luxury perfume bottle on a marble surface, dramatic studio lighting, soft caustic light reflections from the glass, slow rotation revealing all angles, black background, commercial photography aesthetic, ultra-sharp focus throughout
'360-degree orbit' is a clear camera instruction Runway understands. 'Commercial photography aesthetic' sets quality expectations. Specifying 'caustic light reflections' gives Runway a specific visual effect to generate. 'Sharp focus throughout' counteracts the natural depth-of-field bias.
The woman in the image turns her head slowly to look directly at camera, slight smile appearing, hair moving gently in a breeze, eyes tracking to meet the viewer, natural micro-expressions, realistic skin texture, soft natural lighting
Image-to-video works best with subtle, realistic motion rather than dramatic action. 'Micro-expressions' and 'natural' signal the level of realism. Breaking the motion into a sequence (turns head, smile appears, eyes track) gives temporal structure. Hair movement adds believability.
Fluid simulation: luminous cyan ink dissolving into dark water in slow motion, fractal branching patterns, macro photography perspective, tendrils of color expanding outward then curling back, high contrast, black background, loopable motion, dreamlike quality
'Loopable motion' is a useful instruction for motion graphics use cases. Describing the specific fluid behavior (expanding then curling back) creates more controlled motion than 'fluid simulation' alone. Macro perspective concentrates the visual interest.
Browse our full library of Runway prompts.
Browse Runway Prompts