Spotlight: Outfit Finder — turns 'where's that from' into search terms, not links
Everyone has the same small, recurring problem: you see an outfit, on someone's Instagram story, on a stranger on the subway, in an old photo of your own closet, and you have no idea how to find the actual pieces. You end up cropping the image, running it through Google Lens, and squinting at reverse image search results that are mostly unrelated. Outfit Finder is built for exactly that ten-minute rabbit hole.
You upload a photo, and it breaks the outfit down item by item: top, bottoms, outerwear, shoes, bag, accessories. Each item gets its own card with color, material, pattern, and style notes, plus a guess at the brand if there's a visible logo or label. Then, instead of trying to be a shopping engine itself, it gives you two ready-made search queries per item, one aimed at finding the exact piece, one aimed at a similar alternative in case the original is discontinued or you just can't afford it.
The smart decision here is what it doesn't try to do. It doesn't claim to have a product database, doesn't fake a 'buy now' button, and doesn't pretend to always nail the brand. The FAQ is upfront that results are AI suggestions, a starting point, not a guarantee. That's a more honest posture than most 'AI shopping assistant' pitches, which often imply they can match you to an exact SKU. Outfit Finder is really a very good pre-search tool: it does the tedious part (naming what you're even looking at) and leaves the actual purchase decision to you and a search engine.
The other notable choice is where it lives. This isn't a standalone app with its own signup and billing, it's a bot on Poe, which means it inherits Poe's chat interface, its points-based usage system, and its distribution to people already using Poe for other AI tools. That's a low-friction way to launch something narrow like this: no app store, no landing page funnel, just another bot in a platform people already have open. The tradeoff is that you're tied to Poe's UX and pricing model rather than owning the relationship with your user.
Who should try it: anyone who screenshots outfits they like and then does nothing with them, secondhand and thrift shoppers trying to identify pieces they already own, or anyone who wants to recreate a look from a magazine or street style photo without spending twenty minutes reverse-searching each item by hand. It's a genuinely useful five-minute tool for a genuinely annoying five-minute problem.
Who shouldn't bother: if you need reliable brand or SKU-level accuracy (say, for resale authentication or insurance documentation), this isn't built for that and says so itself. If you want an actual shopping experience with real-time prices, stock, and checkout, you'll still end up on Google or a retailer's site doing the last mile yourself, Outfit Finder only gets you to the search bar. And if your photos are dim, cropped weirdly, or shot from far away, expect vaguer results, the FAQ is honest that clear, well-lit images work best.
Where I'd guess this goes next, and this is a guess, not anything the team has said: the obvious next step is skipping the manual search step entirely and pulling live product links or prices from retailers directly, or letting people save an identified outfit to compare later. Whether that happens depends on whether this stays a fun Poe side-bot or turns into something with its own retail partnerships behind it.
Try Outfit Finder: poe.com
See the launch: Outfit Finder on welaunch.sh