ReVibe AI

Title Page
Customizability Options
Gallery

Inspiration

Designing a room, whether its in your house or in an office space, can be overwhelming. There are many options to buy from local furniture shops and online. We wanted an application that reduces the overwhelming amount of brainstorming and furniture options into a realistic design within your desired budget and taste.

What it does

It uses AI to instantly redesign any room based on your style, space, and budget, and generating realistic layouts while linking you to furniture in your area you can actually buy.

How we built it

We built this project using React and TailwindCSS for the frontend and Node.js + Express for the backend. We integrated Gemini AI for our Image and text generation features and SerpAPI for the web scraping functionality in order to bring the user the information and url's to purchase the furniture that was generated.

Challenges we ran into

Our first few hours were spent building the architecture of the app and trying to find API's that would provide furniture for us. Many times we would run into the issue of not having the proper way to provide furniture with our app architecture. Eventually we were able to use Gemini AI and SerpAPI to provide the user with a wide range of furniture options for the type of room they wanted. Our biggest issue, however, was image generation. We had lots of trouble with getting our AI image to generate. We ran into so many errors and have probably memorized all the error codes by now, but eventually, we had solved it by testing different Gemini models until we found the one that worked.

Accomplishments that we're proud of

We are very proud of successfully integrating Gemini AI and SerpAPI and having them communicate with each other in order to generate product links based off the image that the user generated. We are also proud that we were able to properly use Git in order to work on different features in our project at the same time. The biggest achievement that we are proud of is that we had been able to properly build out our application architecture on paper and implement in code.

What we learned

We learned how to effectively use git and divide our work in branches to seamlessly merge our contributions throughout the project. We learned how to effectively integrate multiple APIs and have them communicate seamlessly within one application. We deepened our understanding of working with AI models for both text and image generation, as well as handling and debugging complex API errors.

What's next for ReVibe AI

We want to integrate interactions with users to where they can upload their designs online to other users. We want to add a way for users to store their favorite furniture so they can have easy access to it when they want to remodel their rooms. More customizability options for users to narrow down the design they desire for their rooms.