There are a number of Black creatives who have taken the initiative to document Black fashion - many using social media to catalog and spread awareness of Black culture and it’s impact on society from past to present day: archivealive by Tianni Janae', Black Fashion Fair by Antoine Gregory, and Black Fashion Archive by Rikki Byrd. However the social media accounts aren't optimal to search through, leaving users scrolling through the accounts for inspiration.
Shelby Ivey Christie, a prominent Black fashion historian, discusses record-keeping as a means to right the wrong of American society's erasure of Black contributions for centuries, especially with its export of Black American culture. This in combination with cultural appropriation and the exit of iconic subcultures like GHE20G0TH1K, the need to document and simultaneously protect culture is evident. This digital archive centers around recognizing Black American cultural impact in entertainment, including across the global Black diaspora, and cataloging personal style.
My professional experience is mainly with client side Javascript and React apps, so this was the perfect opportunity to learn more about server side rendering. I wanted to include AI personalization in this app so we're starting with OpenAI and later will compare to Google's Gemini.
For this app to scale, I included authentication with Clerk. They offer quick and easy user management - no more trying to figure out google auth!
I created a local database to quickly get started with the app using prisma and postgresql. Starting with the ORM gave me an idea of how I wanted to structure the data. Once I established that, I decided to switch over and use a headless CRM to pull content from - that way anyone could add data using the Prismic UI at a later point and not have to be an engineer by any means.
Langchain and zod made it pretty easy for me to interact with OpenAI. With our Next.js app, it'll run server side in node. First, I'll use StructuredOutputParser and zod to create a schema of instructions to feed to OpenAI. That will be passed to Langchain's prompt template that will be used to ask the language model for an answer. We then format that answer into the data structure we want!
Want more insight into the product UI and UX? Read more here!
