Clear Cosmetics

• 

  HOME PAGE

• 

  INPUT PAGE

• 

  DATA INPUT

We are so distracted by what cosmetics we wear that we forget what is actually in them.

What it does

Clear Cosmetics is a website that will allow users to check if their products have potentially harmful ingredients in them.

How we built it

Clear Cosmetics was built using Open AI's API with ChatGPT 3.5 as the model. It uses Open AI's API to do the embeddings and then stores it in a Chroma database. It is run using flask as the backend and next.js as the front end.

Challenges we ran into

It was difficult trying to figure out how to get my vector embeddings working. It was also difficult to get the database populated.

Accomplishments that we're proud of

I am super proud that I was able to get the RAG aspect working as it made the whole project what it is.

What we learned

I learned a lot about vector embeddings and how to implement RAG. I learned more about next.js and honed my skills more.

What's next for Clear Cosmetics

I am hoping to add a feature that would just take in a product from the user and then be able to find the ingredients to pass to the model. I would also like to fix the sources as it is having trouble with citing the exact studies and relies more on the actual website I used that corroborated all the data.

Built With

• chat-gpt
• chroma
• flask
• next.js
• openai