Fashionista

Inspiration

The decision fatigue and overwhelming number of choices in online shopping inspired us to create a solution that simplifies the shopping experience and helps users find the perfect outfit quickly.

What it does

Fashionista is an AI-powered chatbot that understands user preferences and curates personalized outfits from various retailers, streamlining the online shopping experience.

How we built it

We utilized OpenAI's GPT-4 for conversation, OpenAI's Ada for embeddings, Pinecone for vector search, Supabase for database, and web scraping for product data extraction. We built a user-friendly UI in the form of a Chrome extension.

Challenges we ran into

Integrating multiple data sources, optimizing search algorithms for product relevance, and finding clever workarounds to scrape the product data were some of the challenges we faced during development.

Accomplishments that we're proud of

Successfully developing an AI-powered chatbot that accurately curates outfit recommendations, addressing decision fatigue in online shopping, and enhancing the overall shopping experience.

What we learned

We gained valuable insights into LLMs and ways to optimize embedding search.

What's next for Fashionista

Next steps for Fashionista include expanding to more retailers, refining the recommendation engine, talking to users for feedback, and exploring direct website integrations.

Built With

• embeddings
• flask
• gpt-4
• javascript
• openai
• pinecone
• python
• supabase