Us | Devpost

Inspiration

The inspiration came from the frustration of traditional AI assistants that provide information in boring, text-heavy formats. We wanted to revolutionize how people consume and interact with information by combining natural language processing with dynamic visualizations and UI elements.

What it does

Our app allows users to ask any question or query through a conversational interface, and it responds by generating beautiful, interactive UI elements tailored to that specific query. Whether it's step-by-step guides, image galleries, data visualizations, or even complex UI flows, our app brings information to life in the most engaging and intuitive way possible.

How we built it

We built the frontend using Next.js, allowing us to create reusable React components for different UI elements. The core logic utilizes Instructor.js to manage the chains of thought - one chain determines the appropriate UI component based on the user's prompt, while another chain gathers relevant data to dynamically populate that component. We used Zod for data validation and parsing throughout this process. This modular, component-driven approach powered by Instructor.js gives us flexibility in generating rich, tailored visual experiences.

Challenges we ran into

Integrating Instructor.js to orchestrate the different chains of thought was a key challenge. Defining the logic flows and handoffs between chains required careful planning. Zod's type-safety also took some adjustment compared to traditional JavaScript object handling. Additionally, optimizing the app's performance when rendering complex components on the fly was an ongoing effort.

Accomplishments that we're proud of

We're proud to have created a truly innovative and visually intelligent AI assistant that redefines how people interact with information. Our app's ability to dynamically generate beautiful UI elements based on conversational input is a groundbreaking achievement that bridges the gap between natural language and user experience design.

What we learned

This project was a serious skills level-up - Instructor.js for complex logic, Zod for data validation, modular front-ends with dynamic pipelines. But the biggest takeaway was just how crucial it is to untangle spaghetti code through proper separation of concerns and clean abstractions. And dipping into cutting-edge tech opened up new realms for pushing UX boundaries

What's next for Us

We're gunning to capture 1% (1.8M) of ChatGPT's 180 million users through aggressive marketing of our visually intelligent interface. From there, convert 1% (18K) of those to paid subscribers at $20 monthly resulting in $360K MRR.

The core focus will be pushing the boundaries of our dynamic UI generation capabilities and further training our AI model to craft tailored, delightful visual experiences on the fly.

Ultimately, our vision is to become the definitive platform for accessing knowledge through an immersive, visually rich lens that leaves text-based interfaces in the past. Continuous AI and UX innovation will drive our quest to revolutionize how the world consumes information.