Inspiration
When my family renovated our bathroom, ther renovations only took 2 days while finding the right vanity, tiles, and bathtubs took about 2-3 weeks. In the end, we also had an extra 6 tiles that are just sitting in our garage now. That is when I realized that there has to be a way for people like my family and other renovators to spend less time and money on product selection, and also save on wastage. This is what made me thing of Agravitaz!
What it does
Agravitaz is an AI powered renovation visualizer platform that compiles home improvement products from multiple retailers into one platform, allowing the user to visualize all the products in their own space, get AI recommendations and information from product specific AI assistants, and a room layout agent, and build, save, and share room states for the opinions of others. Users can also filter through and search for products/brands, and also add products to their wishlist or favorites.
How we built it
We used unity for the the AR components, flutter for the frontend, and rust for the backend. We used openAI for the product specific AI assistants, and are collecting data from users to train a room layout model/assistant. We are also building a retailer research/contact program using AI for admin work and outreach for onboarding new retailers onto the platform.
Challenges we ran into
The AR component, and merging the 3D models into the live AR was difficult and required many XR/AR features we didn't know about before. The phone model also plays a big role in the quality of the AR, so we have to set a few constraints.
Accomplishments that we're proud of
We built a really creative UI for all the working pages like the auth screens, catalog screen, and AR screen. We have a production ready database and have all the cloud services setup for deployment, and we've come this for combining many different tools and services, some of us have never touched in our life before this.
What we learned
Not everything is commercial proof. A lot of the original room layout data nad models we looked into were always restricted from commercial use, so we figured out a way to implement our own data collection system in the application for model training in the future.
What's next for Agravitaz
Deployment, onboarding, successful user marketing, and hopefully some revenue in the near future.
Log in or sign up for Devpost to join the conversation.