Yoink

Inspiration

Remember when you had 10k tabs open just for that one outfit you wanted? We wanted convenience and customization, so we created a web automation tool where you can purchase your fully customized wardrobe to minimize the 10k tabs to just one area, inspired by tinder, we have a quick and convenient interface.

What it does

Our project is automated to access all clothing websites for you, as well as pull exactly what you asked for and present them as options to you with as such as style/fit, color, material, popularity, stock, rating, recommendation percentage and other details. Users are able to swipe right to wishlist, left to discard, and it adapts to the users preferences this way. Users can go to their cart and with the click of a button, payment is completely automated, making purchases more efficient; no need to enter card numbers, phone numbers, emails, since it knows YOU the user!

How we built it

We outlined the plan for storing user, clothing, and store data in the Ghost database, and then provided a way to bridge Tinyfish to the Ghost database since the Ghost database is powered with PostgreSQL. We used Tinyfish to make an API which fetched information from clothing websites with specific details/aspects such as style, cut, making sure it had a certain rating etc. We then used figma and claude design to make the actual project itself and added details/fonts, colors and the logo which is our pig.

Challenges we ran into

The AI Agent was automatically buying things without user preference at first and another challenge was some of the websites were not opening with the link we had for each article of clothing. Additionally we were still deciding on which sponsor platforms would be the most interesting and helpful to use but we narrowed it down it down after talking it out with the sponsors and asking them for input based on our idea.

Accomplishments that we're proud of

We are proud of learning how to use new platforms for the first time and since most of us aren’t the most experienced with agentic tools etc we are proud of being able to learn quickly and tinker with the sponsor tools to successfully implement these tools in our project.

What we learned

We learned how to use tinyfish and ghost in order to create our project successfully and also learned how to manage our time since one of the things we learned was a lot of the steps in making an AI-agent and training the models we used required a good amount of loading time. Additionally, we learned how to manage our time since the agents would take time searching the web etc.

What's next for Yoink

We plan to work on Yoink further by publishing it as a mobile application that will format effectively and efficiently on a phone because most people who shop for clothes casually do so on their phones. Additionally, it felt like the user interface would suit best for mobile applications since most of the user interface involved swiping.