Conjure

Inspiration

The idea for Conjure I actually was inspired by the ideation process for UCI's Designathon this year. In the process of trying to think of a good product to build for the designathon, I thought of many different ideas, but none of them landed and there were so many things that were going on through my head. It was only until I went for a drive and was talking to my friends about the competitino and suddenly the creative juices were flowing. That's when I realized that many of the great ideas we have, not just as designers but as everyday people, come from your ordinary life experiences. Whether you're talking with a friend, eating food, or traveling. Many of these ideas pile up and they start to stack in your brain. Once you go down to sit on your computer, the creativity is lost. I would stare at a blank artboard, unsure of where to even begin, and eventually close the tab without having made a single frame.

I believe that this experience is not unique to me. The gap between having a design idea and being able to act on it is enormous, and it is almost entirely caused by the tools that are supposed to help. Most wireframing tools are built for designers who already know what they are doing. They assume you have a structure in mind before you begin. For everyone else, including product thinkers, students, entrepreneurs, and non-designers with real ideas, the tool itself becomes the barrier. The chaos in your head never makes it to the screen because the path there is too steep. Conjure was my attempt to close that gap entirely. The design-a-thon theme of "Chaos into Clarity" felt, to me, like a direct articulation of the problem I had been living. I wanted to build something that could take the raw, messy, half-formed idea in your head and turn it into something you could actually look at, share, and build on, without requiring you to already know how to design.

What it does

Conjure is an iOS app that turns a spoken description of an app idea into a set of wireframes in seconds. The core interaction is simple: you open the canvas, hold the microphone button, describe what you are imagining, and Conjure generates a structured wireframe of your screens automatically.

Once a wireframe is generated, users can interact with individual screens on the canvas, zoom in, and inspect the structure. The export options are where Conjure becomes a real workflow tool rather than a standalone prototype. Users can send their wireframe directly to Figma as frames or copy a view-only link to share with collaborators. The intention here was to make sure Conjure fits into how designers and product teams already work, rather than asking them to change their process to accommodate a new tool.

How we built it

For research, I used Gemini's deep research function to conduct secondary research on existing voice-to-design tools and the broader landscape of AI-assisted wireframing. This helped me understand what had already been attempted in this space, where prior tools had fallen short, and what users consistently reported as friction points when moving from idea to prototype. That research directly shaped several of the design decisions in Conjure, particularly the emphasis on speed, minimal setup, and seamless export.

I then gathered my synthesized findings from research and started designing originally in lo-fi wireframes using Notability's drawing tools. From then, I created hi-fi designs on Figma and prototyped.

Challenges we ran into

The biggest challenge that I ran into was actually in the beginning, in the ideation phase. I originally had so many ideas, and none of them really resonated with me. It was actually the surplus of ideas that gave light to the project that I came with today.

It was also difficult working with just one person because I had to fill in many roles that I usually would do by myself. Working in this time-pressured environment was quite a struggle for me; however, I am glad that this experience was able to deepen my skills in many different areas.

Accomplishments that I'm proud of

I am proud of how I was able to manage to ideate, research, and synthesize my findings after such a long period of time of frustration. Originally, I had come up with around five original ideas, and all these ideas I was juggling and not confident in. I somehow managed to pull it together and design something that I would actually use in real life, and proud of.

What we learned

Building Conjure taught me that the hardest design problems are not always about the interface. In this case, the most interesting design challenge was in the space between language and structure: how do you take something as open-ended as a spoken thought and translate it into something as structured as a wireframe? I learned that this translation is not a solved problem, and that the quality of the output depends enormously on how well the underlying model understands design conventions, not just language.

What's next for Conjure

The next step for Conjure is to deepen the Figma integration. Sending frames is a good start, but I believe the more powerful version of this feature would preserve component relationships, link screens together with prototype flows, and map generated elements to Figma variables where possible. That would make Conjure a genuine first step in a professional design workflow rather than a tool that requires cleanup before it becomes useful. Longer term, I want to explore what Conjure looks like for teams. Right now it is a single-user tool, but many of the moments where ideas get lost happen in collaborative settings: meetings, brainstorms, client calls. A version of Conjure that could capture a shared conversation and produce a shared wireframe from it feels like a natural extension, and one that I would be excited to build.