Cadence

I am solving the $37B problem of lost project context. Cadence turns team conversations into visual project maps—automatically. AI captures meetings, suggests actions, and updates architecture.

Comment

About the project

The lightbulb moment that wouldn't leave me alone

Back in 2024, I was in another one of those painful product meetings. You know the type, two hours of talking, five minutes of actual decisions, and then we all scrambled to remember what we agreed on. I kept thinking: there has to be a better way.

That's when it hit me: what if AI could just... join our meetings and automatically map out everything we discuss? Not just transcribe, but actually understand our project architecture and keep it updated in real-time.

Learning from the legends

Around the same time, I was obsessing over Andrej Karpathy's talks about how we need entirely new GUIs for the AI era. That sealed it for me—I wasn't just building another meeting tool. I was creating a visual language that both humans AND AI could understand. A project map that would evolve with every conversation.

From napkin sketch to reality (the scrappy way)

Like any good founder story, this started with a notebook. I spent hours sketching rough wireframes of what this "connected knowledge" would look like. Boxes, arrows, sticky notes, the whole messy process.

Then came the grind. Translating those chicken-scratch drawings into actual designs took forever. I leaned heavy on AI tools to speed up the process (hey, if you can't beat 'em, join 'em). Got major inspiration from Dribble's clean UI gallery and tried to make something that didn't look like it was built in someone's garage... even though it basically was.

The hustle never stops

The landing page was pure blood, sweat, and caffeine. I studied 6-7 different high-converting pages, dissected what made them tick, then Frankenstein'd the best parts into something uniquely Cadence. Copy, design, user flow—everything had to nail that "holy shit, I need this" moment.

What I learned (the hard way)

Building this taught me that the best ideas come from your own pain points. Every feature in Cadence exists because me and my friends lived through the frustration it solves. The visual mapping? Because I was tired of losing context. The AI suggestions? Because manual updates suck.

The challenges that almost broke me

Design paralysis is real. I spent way too long perfecting pixels instead of validating the core idea. Copywriting is an art form. Making complex technical concepts sound simple and compelling? Brutal. Scope creep is the enemy. I had to ruthlessly cut features to ship something people could actually use.

Why this matters

Cadence isn't just another productivity tool—it's the GUI for how teams will work with AI in the future. When LLMs can understand our project context as well as we do, that's when the real magic happens.

That's how Cadence was born. One frustrated meeting at a time. 🚀

Built With

Share this project:

Updates