GEN 1 · Anger — 28s · T2Free · copies cast, sets and prompts to your workspace · you pay only when you generate
About this template
Each character starts as a model sheet generated with GPT Image 2.5 Flare at 2K and 4:3, showing front and side views, a short info block and a labeled color palette. Sam's sheet pins down the milky translucent shell, dark metal joints, pale blue eyes and the soft blue glow in the chest, while Del's sheet fixes his hair, goatee, grey tee and leather holster. The interview room is generated separately as an empty plate, then a staging master combines both characters and the room into one wide frame from the same camera position. The video prompts tag every reference (@Image1 for Del, @Image2 for Sam, @Image3 for the room, @Image4 for the staging) and restate the key traits, so the 4 reference image nodes on the canvas anchor every shot.
Each video generation is written as one continuous scene split into numbered shots, with every shot picking up exactly where the previous one stopped: 7 shots across roughly 28 seconds in the first take and 8 shots in the second. A global block at the top sets the look once — raw photoreal footage, a single hard lamp over the table with near-black surroundings, muted low-contrast color with no grading, and a constantly shaky handheld camera. A staging block then locks screen direction, keeping Del at the left end facing right, Sam at the right end facing left and two officers along the back wall. Rendered at 1080p in 16:9, this produces wide masters, a top-down insert of hands on steel and tight close-ups as Sam's fists clench and slowly relax.
Duplicate the template, then regenerate the two character sheets with your own detective, robot, product or creature while keeping the same sheet layout. Replace the room plate with your own setting and rebuild the staging master so everyone sits in fixed positions. Update the character descriptions inside the video prompts but keep the global look block and the numbered-shot structure, which is what holds continuity together. Generate the takes, trim them on the timeline and export to mp4 or FCPXML.