Runway Gen-4.5 image-to-video capabilities explained: from first frame to controllable shots

Use the Runway Gen-4.5 image-to-video pipeline to create a demonstration clip in the style of a "capability explainer": first the pipeline generates a first-frame image with a clear camera position and lighting structure, then a prompt describing motion and camera movement advances the image for 5 seconds, intuitively showing how first-frame anchoring + cinematic language control work. The real inputs for this pipeline are only four types: prompt (motion and camera description, ≤1000 characters), image_urls (a single first frame, jpeg/png/webp), duration (5/8/10 seconds, default 5), aspect_ratio (default 16:9, the first frame will be center-cropped). In other words, the "control" of image-to-video mainly comes from two things: the composition and lighting given by the first frame, and the explicit descriptions in the prompt of camera position, height, and direction of motion. Use cases: early-stage film previsualization and storyboard validation, drafting atmospheric shots for brand and product short videos, and quickly turning static keyframes into reviewable motion clips in design workflows. Developers can treat this example as a minimal viable input template: one first frame + one motion description + two enum parameters for duration and aspect ratio, and similar shots can be reproduced.