Ai PipelineFigma Weavy
A node-based AI creative pipeline that generates photorealistic campaign imagery of any brand's product in cinematic real-world environments — without a single photo shoot. Built end-to-end in Figma Weave, the tree takes a brand logo and outputs fully-styled, publication-ready campaign assets.
BUILT IN
figma weavy
6
2
PHOTO SHOOTS
AI GENERATED
CASE STUDIES
PIPELINE NODES
0
100%
A NEW KIND OF
CREATIVE WORKFLOW
Most brands spend thousands on campaign shoots — location scouting, talent, equipment, post-production. This pipeline eliminates that entirely. By building a structured node tree in Figma Weave, I've created a system that can take any brand's logo and produce cinematic lifestyle campaign imagery in a fraction of the time and cost.
The tree is modular and reusable. Swap the brand logo, adjust the mood board, and the entire pipeline regenerates with a new visual identity. This case study uses Wayders Custom Motorcycles as the primary client example and Losterah as proof the system scales.
Brand-agnostic — feed any logo into the tree and the pipeline adapts the entire visual language around it
Cinematic output — outputs match the quality of a professional lifestyle photo shoot with a human model and location crew
Fully structured — each node has a defined purpose, making the pipeline repeatable, scalable, and easy to hand off
Publication-ready — outputs are sized and styled for social, web, and campaign use without additional post-production
THE PROCESS
PipelineTHE
NODE 1
Mood Board
VISUAL DIRECTION
The mood board node is the creative foundation the entire pipeline builds from. Before any image is generated, reference visuals are fed into the tree to establish the campaign's visual DNA — tone, lighting style, colour temperature, contrast, and overall aesthetic direction. Every node downstream inherits from this single source of truth, which means the final
outputs don't just look good in isolation, they look like they belong to the same campaign. For Wayders, the mood board locked in a raw, road-worn aesthetic — natural light, industrial textures, and the kind of gritty authenticity that defines motorcycle culture. That decision shapes every character, every location, and every shot that follows.
NODE 2
SHOT TYPE
CAMERA DIRECTION
The shot type node replaces the role of a director of photography. Rather than selecting a camera angle from a dropdown, this node operates as a fully structured Art Director system that accepts real cinematography variables — shot angle, lens focal length, aperture, camera body, shot size, and foreground element — and interprets them the way a working cinematographer would. Those variables feed into an Art Director prompt that translates them into rich, precise scene direction a generation model can execute as a single photorealistic frame.
For Wayders, this meant Kodak Portra 400 colour science, a 35mm lens at f/2.8, and a strong low-angle frame that places the subject towering against an open sky. Every variable is a deliberate creative decision, not a default.
NODE 3
Character Generation
TALENT DEVELOPMENT
The character node eliminates casting entirely. Instead of sourcing talent, this node builds a photorealistic human model from the ground up using structured inputs — race, hair type, body type, gender, outfit, and facial hair — which are compiled by an Art Director prompt into a single, production-ready character specification.
That spec then moves through a multi-stage pipeline: a Character Simplifier distills it into a clean generation-ready reference string, a full-body model renders the character head to toe, and a Character Design node produces orthographic front, left side, and back views with no cropping, no missing limbs, and no perspective distortion.
The result is a fully realised, consistent human identity that appears the same way across every image in the campaign gallery.
NODE 4
Location
ENVIRONMENT DESIGN
The location node accepts five inputs — TIME OF DAY, COUNTRY/REGION, SETTING, ATMOSPHERE, and PROP — and passes them through an Art Director prompt that articulates the environment with the precision of a real location scout brief. Light behaviour, surface texture, spatial depth, atmospheric haze, and colour cast are all defined before a single image is generated.
For Wayders, this meant overcast flat diffused daylight over a rural Mediterranean backroad, muted desaturated tones, cracked asphalt stretching to the horizon, with a vintage café racer motorcycle anchoring the midground.
NODE 5
hoodie logo
BRAND INTEGRATION
The hoodie node runs two parallel pipelines — front and back — each generating a blank ghost mannequin garment first before a Compositor node places the brand logo onto the fabric with correct scale, studio lighting, and positioning.
The client's logo files drop straight in, the model renders the hoodie with realistic fabric texture and soft directional shadow, and both composited views feed directly into the Final Output node as the branded garment the character wears throughout the campaign. Swap the logo file, re-run the pipeline, new brand.
NODE 6
FINAL EXPORT
EXPORT & DELIVERY
The final output node is where all five branches converge. A Prompt Concatenator assembles every upstream variable — mood, character, location, garment, and shot type — into a single unified brief, which is then combined with a 10-shot Shot Recipe that defines the full range of campaign variations.
An LLM compiles each recipe entry into a precise, generation-ready prompt and fans it out through Text Iterators across multiple camera bodies, lens specs, and framing combinations — producing ten fully distinct, publication-ready campaign frames from a single pipeline run.
FIGMA WEAVE - FULL PIPELINE VIEW
THE WAYDERS NODE TREE
THE RESULTS
CAMPAIGN
OutputsTwo brands. Same pipeline. Proof that the tree is modular, scalable, and brand-agnostic.
Wayders
CUSTOM MOTERCYCLES
CASE STUDY 01
Wayders is a custom motorcycle brand built around the culture of the open road — raw, mechanical, and unapologetically worn-in. Their visual identity lives in cracked asphalt, weathered chrome, and the kind of dramatic light that only exists at the edge of nowhere.
The pipeline placed the branded hoodie exactly where Wayders belongs — backroads, mountain passes, open terrain — with no shoot, no crew, and no casting call. Just campaign-ready imagery that looks like it cost significantly more than it did.
Losterah
PROOF OF SCALE
CASE STUDY 02
The same pipeline, rerun with the Losterah brand identity. This demonstrates the tree works across entirely different visual languages — proving it's not a one-off result but a repeatable, scalable creative system. Different logo, different mood, different character, same zero-shoot workflow.
SO WHAT?
MattersWHY THIS
This isn't a gimmick. It's a glimpse at how creative production is changing — and a demonstration that I know how to work at the intersection of AI and brand storytelling.
01
A campaign concept that used to take weeks of production planning can be prototyped, approved, and delivered in days. This pipeline compresses the entire creative timeline without sacrificing visual quality.
BRANDS CAN MOVE FASTER
02
SMALL BRANDS GET BIG BRAND ASSETS
Not every company has the budget for a professional lifestyle shoot. This pipeline democratises cinematic campaign imagery — a Wayders-quality result at a fraction of the traditional cost.
03
AI IS A CREATIVE TOOL NOT A REPLACEMENT
The pipeline still requires creative direction, brand understanding, and visual instinct to work well. The AI executes — but a human has to know what to ask for. That's the skill.
04
A REPEATABLE SYSTEM. NOT JUST A ONE-OFF
The Losterah case study proves the tree scales. Any brand with a logo and a visual identity can plug into this pipeline and receive campaign-ready outputs. The system is the product.
WANT THIS
FOR YOUR BRAND?
If you're a brand, agency, or creative team exploring AI-assisted production — let's talk about what this pipeline could do for your next campaign.