Curator.ai
Co-Founder | MIT AI Hackathon Winner | MIT AI Conference NYC |
Forbes Mention | IEEE Tech Forum @ Microsoft
Vison Language Understanding for Contextual Product Recommendation
Curator.AI helps you find the perfect furniture for your space using AI and Augmented Reality (AR). It’s designed to make furnishing easy, fast, and stress-free.
It offers personalized recommendations by understanding your style, space, and budget(not limited to). One can discover cohesive furniture sets, visualize them in your room using AR, and reduce endless browsing time ensuring everything fits perfectly—both visually and functionally. Whether you’re a busy professional, a first-time homeowner, or just overwhelmed by choices, this innovative blend of Vision-Language Models, ARKit/AR Core, RoomPlan API and Google Search API, will keep you stress free.
It’s not just shopping—it’s smarter, waste saving, more intuitive design.
User Stories
Problems
Innovation
"I would likely use this solution because it helps a lot with the search exhaustion”" — Judge's comment
Our AI system begins with 3D object recognition to map out the room's spatial arrangement. This allows us to identify where existing items, such as tables or windows, are their dimensions. With this spatial understanding, we can automatically place products in a contextually appropriate manner, like positioning a sofa in front of a window or determining the position and number of chairs for an existing table in the space. Next, we consider the room’s style and ambiance by using a vision-language model. This helps us identify the room's character—whether it's a cozy, traditional living room or a sleek, modern office. Based on this contextual analysis, we conduct an internet-wide search to find physical products you can actually purchase. We are not generating new 3D objects, we recommend real, purchasable item that works with your space.
And this is displayed in AR. Because we want you to see how it would actually look like in space. How it fits and how it feels. Since we are in AR, we can’t type, therefore we we are leveraging speech and generative voice AI. Where you can use your conversation to filter things with budget, rearrange objects or find new items.
This sequence diagram shows Curator. Ai's two main flows: Vision-Language Product Recommendation and Generative Voice Interaction highlighting -multiple technologies for personalized and interactive AI-driven recommendation. Vision language AI flow gives a contextual understanding of the room. And the Generative Voice AI flow gives us the user’s intention. So, When the user initiates, these two data flows work together to create a contextual product recommendation which is fed into our data retrieval algorithm to search real products. The system also includes error handling, prompting the user if context is unclear or transcription fails.
Speech + Vision into products
"The product solves a very unique problem that many of us come across.”" — Judge's comment
Market Size (USA) in 2023
"Having AI provides the appropriate suggestions and ability to visualize the product makes it a good fit for the user”" — Judge's comment