TechStyle

🛍️ Inspiration

Imagine a world where shopping online for fashion isn't just about searching for a brand name or a specific item, but about discovering your style and personality effortlessly. Discover e-commerce in a new way with our AI-driven solution! We're here to make shopping a joyful and curated experience, where every individual finds not just clothes but a reflection of their identity.

The Problem

When shopping online, customers often struggle to find the right terms to describe what they want, turning the search into a frustrating task. In 2022, 58% of global brand searches were unbranded, revealing a huge opportunity to capture sales. Even so, many users browse anonymously, making it difficult for retailers to personalize the shopping experience effectively.

What it does

🔍 AI-Powered Search Filter

• Users can try the public interface to seamlessly find directly linked products from favorite retailers inspired by their unique aesthetic preferences.
• The functionality delves into non-traditional filters, letting users discover items based on vibes, emotions, and styles—think "cozy and relaxed" or "vintage chic"—capturing the essence of what they're looking for.

🛠️ Sandbox & Integration

• Retailers have access to a sandbox environment for testing and refining the service. This allows them to manually import data and see how the AI adapts to their catalog.
• Our APIs offer seamless integration, connecting retailers' existing filtering tags, image catalogs, and product descriptions to enhance search relevance and user experience.

How we built it

We use Cohere and OpenAI for image-to-text conversion and advanced text-based semantic search based on given filters of a company - find the closest combination of filters. The backend and frontend are run on NextJS with Supabase dealing with postgres database and S3 bucket interfaces for storing large data regarding product catalog.

Challenges we ran into

We encountered challenges in prompt engineering and fine-tuning AI requests. With our diverse backgrounds and skill sets, it required effort to align our ideas and develop both the public and business components of the application.

What we learned

Each of us picked up new skills along the way and weren't afraid to pivot when things didn’t go as planned. In the end, we built a working project and are eager to keep exploring the tech space.

What's next for TechStyle

Image Search: Introducing image search functionality to allow users to upload photos and find similar products. Improved Personalization: Enhancing our AI algorithms to deliver more precise recommendations tailored to user preferences. UI Enhancements: Refining the user interface to make navigation smoother and more intuitive. Enhanced Public UX: Focusing on optimizing the public interface for a seamless and enjoyable user journey.

Built With

• amazon-web-services
• cohere
• nextjs
• openai
• postgresql
• supabase