Inspiration
We noticed how the beauty industry exploits women by overcharging them for makeup products. We wanted to make a change to this by helping them find alternatives.
What it does
LookLens lets any woman point her camera at a makeup look she loves and instantly knows exactly what products are being used. It then finds the most affordable dupes within her budget, shows her instructions on how to apply the look, and uses AI to preview what that makeup will look like on her own face. All this without her spending any money to try out makeups.
How we built it
We built it by using shadcn/ui. We built the frontend using Next.js and Tailwind CSS. For the AI, we used Gemma 4 as the main researching agent and Gemini 3 Flash as a adversarial reviewer to help improve correctness in the agent's outputs. We used Nano Banana to apply the identified makeup onto the user's own photo for a personalized preview.
Challenges we ran into
We ran into time management issues and refining the processing so it becomes better.
Accomplishments that we're proud of
Within the time limit, we were able to build a reasonable functional prototype.
What we learned
We learned that building a prototype in a limited amount of time with teammates at different skill levels isn't easy. We need to be more prepared next time.
What's next for LookLens
We are going to implement a shop where the user can buy straight from the website. Implement changing settings for the reference photo and your budget after onboarding.
Built With
- clerk
- drizzle
- geminiflash
- gemma4
- nanobanana
- nextjs
- postgresql
- tailwindcss
- typescript
Log in or sign up for Devpost to join the conversation.