🎬 Creative Ad · LTX 2.3 Ingredient

ROOFTOP RUN one sheet, one scene

Four ingredients on a single reference sheet — a coral sneaker, a runner, a rooftop at dusk, a duffel — composited into one cinematic ad scene that keeps each one on-model. Not generic text-to-video: the sheet says what appears, the prompt says what happens. Real assets, end-to-end on Livepeer's ltx-q-ingredient cap.

model
LTX 2.3 Ingredient
cap
ltx-q-ingredient
input
1 reference sheet
render
~2 min · $0.056/s
other tools
gpt-image
The point: identity, not just motion. Generic text-to-video invents a new sneaker and a new person every run. LTX 2.3 Ingredient takes a reference sheet of the exact things you want in the shot and composites those into the scene — the video analog of moodboard-to-scene. Build the sheet, direct the action, get an ad where the product is the product.

The two halves — ingredients in, scene out

Left: the reference sheet (built with gpt-image) — four labeled cells, neutral backgrounds, consistent scale. Right: the generated scene from ltx-q-ingredient with the action prompt. The sheet locks identity; the prompt directs the shot.

Reference sheet: coral sneaker, runner in teal, rooftop at dusk, charcoal duffel
INPUT · reference sheet (gpt-image)
OUTPUT · ltx-q-ingredient scene

Prompt: "the runner in the teal windbreaker laces up the coral sneaker on the city rooftop at golden-hour dusk, the gym duffel beside them; slow push-in as the sneaker catches the last warm light." The sneaker, the jacket, the rooftop, and the bag all carry through from the sheet.

How it was made — the recipe

01 · SHEET
gpt-image · the ingredients
A 2×2 labeled grid — SHOE / RUNNER / SETTING / PROP — on neutral backgrounds, consistent scale, so each ingredient is unambiguous. (Or assemble real product photos with pillow-grid.)
02 · SCENE
ltx-q-ingredient
create_media(action:"animate", source_url:<sheet>, model_override:"ltx-q-ingredient", prompt:<action + style>). The sheet is the source; the prompt is the direction.
03 · MORE SCENES
same sheet, new prompts
Re-run with different action prompts on the same sheet for a multi-shot ad — the ingredients stay consistent across cuts.
04 · FINISH
sonilo-v2m · ffmpeg
Auto-score with sonilo-v2m (or music), ffmpeg-concat multiple scenes, ffmpeg-export to aspect, ffmpeg-overlay a logo.

Verdict: ships. One reference sheet → a coherent, on-model ad scene in ~2 minutes, where the sneaker, character, environment, and prop all survive into the video. The identity-preservation that generic text-to-video can't give you.

Pair it with the reference-sheet-to-ad playbook and the ltx-ingredient skill for the full recipe + reference-sheet craft.