37spots

Discover awesome brands. Shop application based on AI

Comment

🧠 Inspiration

Web search is rapidly shifting away from traditional engines like Google and toward AI-powered experiences. In the realm of e-commerce search, even the biggest platforms often fail to deliver optimal user experiences. Given these shortcomings, now is the right moment to reimagine how people search for products online — and do it better.

🌍 The Global Shopping Directory Concept

The dream of a global shopping directory — the holy grail of e-commerce — has been around for years. Many have tried to build it. Many have failed. Why?

Because of unstructured data.

Take any product you can buy online: on the brand’s website, you'll find the product name, description, specs, price, and maybe reviews — but all in different formats, layouts, and structures.

Extracting and unifying this data across thousands of brands is nearly impossible using traditional scrapers. Each website is unique. Parsing HTML breaks quickly.

That’s where AI — especially Perplexity Sonar — changes the game.

⚙️ How It’s Made

We start with just a photo and a few spotted products, each linked by a raw URL.

Here’s where the magic begins:

Using the Perplexity Sonar API, we extract:

  • Product name, price, and currency
  • Product description
  • Structured product details (in JSON)

37spots is a platform where users upload photos and manually spot products (via URLs). At this stage, the application doesn’t “know” what’s in the photo.

But Sonar bridges that gap.

Instead of relying on brittle scraping rules or manual data entry, Sonar analyzes the raw URL and retrieves accurate, structured information directly from the product page — no matter how the site is built or styled. It understands the content semantically and outputs clean, uniform data in real time.

This means:

  • No more writing custom parsers for every website
  • No dependency on layout or HTML structure
  • Scalable extraction across thousands of brands
  • Structured product data you can use instantly

It’s a powerful leap toward building a truly global, searchable shopping directory — all starting from a photo and a few smart API calls.

🔮 What’s Next for 37spots

  • Launch 37 unique types of interactive spots to enrich shopping photos (text, video, polling, geo and so on)
  • Introduce the 37spots token (total supply: 37 million) to power a new app ecosystem
  • Build curated product collections using structured product data extracted by Perplexity Sonar
  • Implement AI-powered product search driven by natural language prompts and Perplexity Sonar
  • Develop advanced AI recognition models for categories like plus-size clothing, lookbook styles, room types, and more
  • Create an embeddable shoppable photo widget for seamless integration across websites
Share this project:

Updates

posted an update —

37spots.com: The Future of E-Commerce Search Perplexity Hackathon Submission

1) What is the Project? Description: 37spots.com is an AI-driven global shopping directory that reimagines e-commerce search. It allows users and influencers to upload photos, “spot” shoppable products, and discover detailed product and brand information automatically extracted using advanced AI tools like Perplexity Sonar. This system turns unstructured product data across thousands of websites into a unified, searchable catalog.

Technologies:

Frontend: React, Vite, Tailwind CSS Backend: Node.js, Express, PostgreSQL AI & APIs: Perplexity Sonar API, Clarifai, embedding vectors, image analysis models Infrastructure: Docker, Redis, HAProxy, Cloud Storage Key Features: ✅ Upload and spot products in photos ✅ Automatic extraction of product details (name, price, description, features) ✅ Advanced photo analysis (objects, faces, colors, concepts) ✅ Rich brand profiles with warranty, sustainability, and shipping info ✅ Flexible, conversational search with advanced filters ✅ Influencer platform with affiliate tracking (in progress) ✅ Dynamic collections for intuitive browsing

Demo/Preview: Visit Live Demo https://37spots.com Screenshots can be added here once available

2)

Why Does This Project Exist? Motivation: Traditional e-commerce search engines and even big marketplaces struggle to unify product discovery across the global web. With unstructured, scattered data, users face frustration in finding the exact products they want. 37spots leverages modern AI to solve this long-standing problem.

Use Cases:

Shoppers looking to discover products by photo or aesthetic Influencers tagging products in their content Brands seeking a unified discovery platform Researchers analyzing global product trends Goals:

Create a seamless, AI-powered product discovery engine Build the world’s richest brand and product directory Enable new monetization paths for influencers Lay groundwork for future blockchain integration (affiliate payments, ownership)

3) Who is the Target Audience? Primary Users:

Consumers interested in visual, photo-driven shopping Influencers wanting to monetize content without personal sites Brands wanting exposure and fresh discovery channels Developers and researchers curious about AI-driven search

4) Contributors Currently a solo project, but open to:

Developers (Node.js, React, AI integrations) Designers (UI/UX) Data scientists (recommendation systems, embeddings) Documentation writers Assumptions: Some familiarity with modern web stacks (Node.js, React) is useful if contributing.

Thank you for checking out 37spots!

Log in or sign up for Devpost to join the conversation.