AI Product × Fashion Tech
Where it started
A fashion designer told me she'd just paid $5k for a photoshoot. The sample wasn't even confirmed for the final collection yet. The photos came back three weeks later. By then the design had changed.
I talked to more designers. Same story every time. They needed visuals before the sample existed — for buyers, for IG, for pitch decks. The shoot timeline never lined up with the design timeline. So they were either spending money they didn't have or shipping with no visuals at all.
That's the gap Nendo fills. Not the shoot — the bit before it. The expensive, three-week wait where you need a garment on a body, and there's no body, no sample, no time.
What I built
The brief I gave myself: flat-lay photo in, finished lookbook out. No prompts. No subscription. No need to understand a single thing about the AI doing the work.
You upload your design. You pick a model muse — or save one you've used before. You choose editorial or streetwear. You hit generate. Eight looks and a 360° body turnaround come back in under four minutes. High-res ZIP. Done.
No subscriptions either — fashion ships in drops, not monthly cycles, so I built credit packs instead. Buy when you need it. That decision alone cut churn conversations in half.
What the product generates
Same muse, same garment —
4 of 8 editorial poses shown
Silhouette review
A design decision I'm proud of
Early users kept asking: "can I see it in streetwear?" Not just cleaner studio shots — a completely different context, different energy, different buyer.
So I built styles as a first-class concept in the product. Same muse, same garment, same brief — but editorial gives you clean studio energy, and streetwear gives you kinetic poses, a wide lens, and a city backdrop. One extra click. Completely different lookbook. The garment reads differently in each.
That feature alone changed what designers could pitch and to who.
The hardest part
The generation pipeline is genuinely complex — image conditioning, pose estimation, IP-adapters, ControlNet. None of that should exist from the user's point of view.
The hardest design work on this project was deciding what not to show. No sliders. No model parameters. No "try prompt" field. Just: describe the garment, pick the muse, hit go.
Every time I added a control, I asked: does the designer actually need this, or am I just showing off what the tech can do? Most of the time, I cut it.
The Tweak button — "open the jacket", "shorter hem", one-look refinement without redoing the whole grid — took three iterations to get right. That's the one I'm most proud of.
The best way to understand what Nendo does is to use it. The product is the case study.