Case study — AI creative suite

A production suite for AI campaign visuals

Character-consistent storyboards, guided product relighting, and platform-ready campaign assets — organized as a reviewable three-pass production pipeline.

Year: 2025–2026
Role: Product · pipeline design
Status: In production
Stack: Next.js · Vision-LLM director pass · Multi-model image routing · Zod · Mutex-locked storage
Relationship: Selected delivery work; the public case uses a functional project name.
Market: Creative-production workflow; client identity is not disclosed publicly.

The problem

AI images are cheap. Consistent ones aren't.

Any tool can generate one good frame. A campaign needs twenty — same character, wardrobe, and world across the approved shot and format plan format. Prompt-by-prompt generation collapses into a lottery, and marketing teams burn days re-rolling.

The brief: make AI asset production behave like a production pipeline — deterministic passes, reference discipline, and outputs a brand team can actually ship.

The system

How it fits together.

simplified public map of implemented components and boundaries

The output

Compare, frame by frame.

Drag the sliders to compare relight passes, then scroll the storyboard — each frame was rendered in parallel from one anchor image and the director-pass text beside it.

flat studio light

cinematic warm key

Product relight · set 01masked relight pass · drag to compare

daylight neutral

night interior

Product relight · set 02masked relight pass · drag to compare

storyboard run — director-pass text with the frames it produced, one visual DNA across every shot

frame 01
01Wide establishing — the courier crosses an empty dawn street, city half-lit behind her.
frame 02
02Close on the package in her hands; the brand mark catches the first light.
frame 03
03She checks the route on her phone — same face, same jacket, new angle.
frame 04
04Over-the-shoulder: the doorway ahead, warm light spilling out.
frame 05
05The handoff — two hands, the package, shallow focus.
frame 06
06End card: product on the doorstep, morning sun, logo lockup.

The build

Three passes beat a thousand prompts.

Pass 0 — the anchor frame

A run starts by generating one master visual from the user's character and prop references. That frame becomes the visual DNA every subsequent shot inherits — lighting, wardrobe, palette, world.

reference lock

Pass 1 — a director, not a prompt list

A vision-capable LLM reads the script and the reference images, then writes a complete cinematic paragraph per shot — camera, action, continuity. No keyword soup; language a cinematographer would accept.

vision LLM · structured output

Pass 2 — planned frames in parallel

Planned shots can render in parallel, each conditioned on the anchor frame plus its own director paragraph and any character references it needs. Parallel frames share the approved visual direction and remain reviewable.

parallel edit-model renders

Relighting with painted masks

For product shots, users paint a mask over the region to relight; the edit model transforms only that region under adjustable guidance — cinematic environments without touching the product's geometry.

masked edits · guidance control

Files that survive parallel writes

Serverless routes writing shared JSON during 20-way parallel renders corrupt state without discipline — a memory-lock mutex queue serializes writes so concurrent runs do not overwrite each other.

mutex-locked storage

Implementation evidence

Three-pass flowreference, direction, and render stages remain inspectable

Human reviewcreative approval remains a deliberate part of the workflow

Format routingoutput requirements are handled by the pipeline, not prompt memory

Constraint

Teams needed repeatable characters and product treatments across multiple output formats.

Tradeoff

The workflow introduces explicit direction and review passes rather than optimizing for one-click generation.

Public images demonstrate the pipeline stages. Model counts, output volume, and campaign performance are not presented as client results.

For your business

Discuss a related system.

If your team produces campaign visuals, product imagery, or social content at volume, this pipeline pattern turns AI generation from a toy into a production tool — consistent characters, repeatable looks, and formats cut per platform. We design the pipeline around your brand, not the other way around.

Start a project More work

Next caseA community platform with an integrated finance layer →