Inspiration

Shopping for clothes is frustrating either way. Going to stores takes hours you don’t have, and shopping online means guessing whether something will actually fit, which often ends in a return. We wanted the convenience of online shopping with the confidence of a fitting room. That led to Mirra: an AI stylist that puts outfits together for you and shows them on a 3D model of your own body before you buy.

What it does

Mirra is an AI stylist that shows you how clothes will look on your body before you buy them.

  • Your own 3D avatar: Enter your body measurements and Mirra builds a 3D model with your proportions, so you see how clothes fit you instead of a generic mannequin.

  • Styling through conversation: Tell Mirra what you need, like “something casual for a first date” or “an outfit for a job interview,” by typing in the chat or speaking out loud. It puts together a complete outfit for you.

  • See it on you instantly: The outfit appears on your avatar right away, so you can judge the fit and the look together.

  • Know what you’re wearing: A description panel breaks down each piece of the outfit.

  • Shop the look: Mirra gives you a link to buy each item, so you can go from idea to purchase without searching through hundreds of products.

How we built it

  • Next.js powers the whole app, from the frontend interface to the API routes that connect our services. Google Gemini API is Mirra’s styling brain. When a user describes what they need, Gemini interprets the request, puts together a complete outfit, and writes the description of each piece.

  • ElevenLabs runs our voice assistant, so users can talk to Mirra in plain spoken English and hear it respond, making the experience hands-free.

  • Supabase handles user authentication, so each user has a secure account tied to their own avatar and measurements.

Challenges we ran into

  • Challenges we ran into: Trying lidar scanning, finding it too hard to get working in the time you had, and switching to measurements is exactly the kind of story judges like to see. It shows you made a smart call under pressure.

  • What’s next for Mirra: Lidar scanning from your phone is a natural next feature, because it would make the avatar even more accurate with less effort from the user.

Accomplishments that we're proud of

  • Getting the AI integrated into our system
  • Building an interface that is intuitive and frictionless for users to experience.

What we learned

Our biggest challenge was scope. We set out to build lidar body scanning, so users could scan themselves with their phone and get an exact 3D model. As we went along, we realized that getting accurate scans to work reliably was far too much for the time we had. Rather than end up with a half-working feature, we adapted. We switched to a measurement-based avatar, where users enter their body measurements, and Mirra builds a 3D model with their proportions. That kept the core idea, seeing clothes on your own body, while giving us something we could make work well within the deadline.

What's next for Mirra

Integrating LIDAR scan functionality.

Built With

  • elevenlab
  • next.js
  • python
  • supabase
Share this project:

Updates

Submission history