Inspiration

What inspired me was the long, difficult, and grueling process of traditional interior design. Finding the right furniture, visualizing how pieces work together, and making countless decisions can be overwhelming and time-consuming.

What it does

Give it an image of your empty room, and it transforms your space automatically. The agent uses AI tools to find real furniture on Google Shopping through Arcade, leverages LlamaIndex for intelligent design decisions, generates stunning visualizations with GPT-4.1-mini, and even narrates the final design using Inworld's text-to-speech technology.

How we built it

We built this using a modern tech stack: React with Vite for the frontend, FastAPI for the backend, and integrated cutting-edge AI tools including OpenAI's GPT-4.1-mini, Arcade's Google Shopping API, LlamaIndex for knowledge management, and Inworld AI for voice synthesis.

Challenges we ran into

Deployment proved challenging, particularly dealing with CORS issues between our frontend and backend services. We also faced complexities in coordinating multiple AI services and ensuring the autonomous agent didn't get stuck in infinite loops while searching for furniture.

Accomplishments that we're proud of

We're proud that our system actually preserves the original room's architecture and characteristics while transforming it. The agent successfully creates cohesive, beautiful designs that respect the existing space while adding carefully selected furniture pieces.

What we learned

We gained valuable experience with Railway deployment, discovered the power of new AI tools like Arcade for real-world shopping integration, and learned how to orchestrate multiple AI services into a seamless autonomous workflow.

What's next for Interior Design Agent

Future enhancements include integrating more shopping tools and retailers, implementing smarter search algorithms, upgrading to more advanced image generation models, and adding interactive 3D room visualization for an even more immersive design experience.

Built With

Share this project:

Updates