Tried to promote our lingerie store on Pinterest with no ad budget, ended up building a tool

Hi everyone! Disclosure first: I build Pinlace. It generates on-model product photos for lingerie shops and posts them to Pinterest. I’m here for feedback, not signups.

How we ended up building it

We set out to launch our own lingerie store. The ad budget was small enough that paid acquisition was never realistically the plan, so we went looking for organic channels instead. Pinterest kept coming up - it’s one of the few places where a product image still finds people months after you post it.

Starting from zero was the hard part. Pinterest wants a fresh image close to every day, and a new brand has no library to draw on. A catalog shoot gives you maybe twenty usable images, and booking one isn’t a small line item when you haven’t sold anything yet.

So we tried generating images instead. That turned out to be much harder than expected:

  • most general AI tools block lingerie outright, or give you something you can’t publish
  • keeping the actual product recognisable- the real lace, the real cut - is a completely different problem from generating a nice-looking image
  • getting variety in poses, angles and scenes, so that every pin doesn’t look like the same photo, is its own thing
  • results depend heavily on how you prompt, which isn’t something a store owner should have to learn

We spent long enough on this that it stopped being a side task and became the product.

What it does now

It reads your catalog, generates on-model photos of your own products, and schedules a pin a day with SEO titles.

Worth saying: if you already have photos, you don’t need the generation side at all. We scan your store, pull your existing product images, and you can queue and post those to Pinterest straight away. Plenty of shops just need the posting half.

Limits, so nobody’s surprised

  • Generated images are AI. We think shops should say so, and we’d treat them as lifestyle and mood pins rather than as a substitute for showing the actual garment on the product page.
  • It’s built specifically for lingerie. If you sell something else it will underperform.
  • Pinterest is a slow channel by nature. This is months, not weeks.

Happy to answer anything about how the generation works under the hood - and if you try it and the output isn’t good, tell me. I’ll tune it with you. That’s more useful to me right now than a signup.

Hi Steve,
Thanks - that’s the right question, and the honest answer is that most of it comes down to the prompt.

That’s the part we took on rather than leaving it to the merchant. Our prompts are written specifically to carry a garment’s details across - lace, cut, strap placement, colour - instead of letting the model reinterpret them. Getting that right took considerably longer than building the rest of the app, and it’s why a general-purpose tool hands you something that looks like lingerie but isn’t your lingerie.

The feature doing most of the work here: you can put your actual product onto a photo you already have. Your garment, your existing shot, and the prompt is built to transfer the product faithfully onto it rather than generate a new one from a description.

What I won’t claim: this minimises the problem, it doesn’t remove it. A share of generations will still come out wrong - that’s the nature of the technology, and anyone telling you otherwise is selling.

Fit we don’t claim at all. A generated model wearing something tells you nothing about how it sits on a real body.

On existing photos - that’s exactly who that path is for. Plenty of shops have decent photography and only need the Pinterest half.