Show HN: I Built an AI That Matches Perfumes with Their Closest Twins | Hacker News Reader

Show HN: I Built an AI That Matches Perfumes with Their Closest Twins

dupefinderai.com

5 pointsNabilChiheb4 comments

Hey HN!

Ever wondered if there’s a fragrance out there that smells just like your favorite but from a completely different brand?

That’s what got me thinking, and after a weekend of deep work, I created DupeFinderAI.

What It Does:

DupeFinderAI is an AI-powered tool that scans and compares thousands of perfumes, identifying those that share nearly identical notes and aromas.

Whether you're curious about scent similarities or just love discovering new perfumes, this tool has you covered.

How I Built It:

- Scraped data from 40,000+ perfumes over a weekend - Used that data to train the AI on recognizing fragrance profiles - Focused on making the experience smooth and intuitive

Why It’s Cool:

Finds "scent twins" across different brands Helps you explore new fragrances based on ones you already love Could even be useful for perfumers and fragrance enthusiasts alike

I’d love to hear your feedback, ideas, and any thoughts on how to improve it! Check it out and let me know what you think.

Open articleView on HN

This would be a wonderful tool to package for companies such as ULTA and SEPHORA to add as a widget to their website for finding.

Another item to add would be to scrape the ingredients of (well, anything with fragrance chemicals) for allergens that one might want to filter out. Or specific chemicals to avoid.

Also - can you please share youre scraping method/script/setup? I have a lot of things that I want to do with scraping product functions/features like this - but havent figured out a good way to learn to scrape well (attempting to use Crawlee, for example)

Alos, this would be well applied to Wines.

Thanks for the fantastic suggestions! Packaging this tool as a widget for ULTA, SEPHORA, or similar companies is definitely something I see as a great fit. It could really enhance the user experience by helping customers find products more efficiently.

The idea of scraping fragrance ingredients for allergens is spot-on and could be a game-changer for those with sensitivities or preferences. I’ll explore how we can integrate a feature like that into the tool!

As for the scraping method, I’d be happy to share my approach! I primarily use python/brightdata/Selenium, and it’s been effective for scarping all these informations.

I can walk you through the setup if you’re interested or point you toward some resources to help you get started with it.

Crawlee is a solid choice, too, but I’d be happy to share some tips that might make things smoother for you!

That would be WONDERFUL if you could walk me through -- I'm a solo/lone person and I have found the limitations of learning from AI bots, specifically python - which I am weak with.

One of my best friends worked at The French Laundry for years, and has saved all the nightly menus from working there, and that would be an interesting dataset to train for flavors, notes, wine-pairings... (She now works for one of the most exclusive wineries ($1,000+/btl) one could start making olfactory pairings across all the way that sense influences an experiential environment)... using the logic of how important smell/fragrance is to eating/drinking/life - using the foundation for fragrance similarities, with food/wine/flavor pairings - and for allergen logic.

---

sstave proton me

By the way, the API is now available on RapidAPI

https://rapidapi.com/remote-skills-remote-skills-default/api...

Reply on news.ycombinator.com