copalm
Get Started
Get Started
introducing a knowledge library for everything you've figured out.
introducing a knowledge library for everything you've figured out.
ask copalm to build a library...
Attach
Attach
copalm is a conversational AI platform to document your expertise.
copalm is a conversational AI platform to document your expertise.
It gives experts a place to capture their knowledge, shape it, and make it usable for others. Not a course builder or a wiki, but a knowledge library that works the way expertise actually lives — layered, contextual, and conversational.
It gives experts a place to capture their knowledge, shape it, and make it usable for others. Not a course builder or a wiki, but a knowledge library that works the way expertise actually lives — layered, contextual, and conversational.
53+ early users on the waitlist.
53+ early users on the waitlist.
23+ SME interviews conducted.
23+ SME interviews conducted.
513+ design and engineering iterations.
513+ design and engineering iterations.
from capstone to company.
from capstone to company.
What started as a graduate capstone project at Indiana University in January 2025 became a registered company in July 2025. That shift changed everything —
What started as a graduate capstone project at Indiana University in January 2025 became a registered company in July 2025. That shift changed everything —
the stakes, the questions, the standard for what good enough meant.
the stakes, the questions, the standard for what good enough meant.
I went from designing for a grade to designing for real people with real problems. That reframe shaped every decision that followed.
I went from designing for a grade to designing for real people with real problems. That reframe shaped every decision that followed.
talked to 23+ subject-matter experts to find that one insight.
talked to 23+ subject-matter experts to find that one insight.
From academics, consultants, instructional designers, independent educators in five months. The goal wasn't to validate an idea. It was to understand how expertise actually gets shared, and where it breaks down.
From academics, consultants, instructional designers, independent educators in five months. The goal wasn't to validate an idea. It was to understand how expertise actually gets shared, and where it breaks down.
Three things kept coming up:
Three things kept coming up:
Knowledge lives in people's heads, not in documents.
Knowledge lives in people's heads, not in documents.
The tools that exist ask you to organize before you're ready.
The tools that exist ask you to organize before you're ready.
The jump from "I know this" to "someone can learn this" is where things falls apart.
The jump from "I know this" to "someone can learn this" is where things falls apart.
These weren't edge cases. They were the pattern. And they pointed somewhere different than where the product had started.
These weren't edge cases. They were the pattern. And they pointed somewhere different than where the product had started.
two pivots that changed the scope from courses to documentation.
two pivots that changed the scope from courses to documentation.
copalm launched with a clear focus: turn expertise into courses. I built toward that, tested it, and watched people use it in ways I hadn't designed for.
copalm launched with a clear focus: turn expertise into courses. I built toward that, tested it, and watched people use it in ways I hadn't designed for.
They weren't building courses. They were trying to capture how they thought, their process, their frameworks, the things they'd figured out over years of doing the work. The course was never the point. The knowledge was.
They weren't building courses. They were trying to capture how they thought, their process, their frameworks, the things they'd figured out over years of doing the work. The course was never the point. The knowledge was.
Two pivots over five months moved the product from course builder to knowledge library. Both came from research & made the product harder to explain and more honest about what it actually solved.
Two pivots over five months moved the product from course builder to knowledge library. Both came from research & made the product harder to explain and more honest about what it actually solved.
designed entirely in code, slightly in figma.
designed entirely in code, slightly in figma.
Figma took a backseat in this project. Every design decision was made directly in Lovable — building, breaking, and iterating in the browser. In GCP I built the architecture that informed the contextual layer of the app.
Figma took a backseat in this project. Every design decision was made directly in Lovable — building, breaking, and iterating in the browser. In GCP I built the architecture that informed the contextual layer of the app.
This wasn't a constraint. It was a choice. Designing in code meant the gap between "what I'm imagining" and "what exists" collapsed almost entirely. I could feel how an interaction worked, not just see how it looked.
This wasn't a constraint. It was a choice. Designing in code meant the gap between "what I'm imagining" and "what exists" collapsed almost entirely. I could feel how an interaction worked, not just see how it looked.
the enlightening chemistry of building while learning.
the enlightening chemistry of building while learning.
For most of this project, research and building were happening at the same time. An interview on Tuesday would change what I was shipping on Thursday.
For most of this project, research and building were happening at the same time. An interview on Tuesday would change what I was shipping on Thursday.
That chemistry or rather tension was uncomfortable and I think somewhere necessary. It meant the product was never too precious to change. It also meant some things got built two or three times before they were right.
That chemistry or rather tension was uncomfortable and I think somewhere necessary. It meant the product was never too precious to change. It also meant some things got built two or three times before they were right.
end to end work at its finest.
end to end work at its finest.
This project is the clearest example of what end-to-end product design looks like when one person holds the whole thing. Research informed architecture. Architecture shaped interaction. Interaction revealed gaps that sent me back to research.
This project is the clearest example of what end-to-end product design looks like when one person holds the whole thing. Research informed architecture. Architecture shaped interaction. Interaction revealed gaps that sent me back to research.
The work doesn't separate cleanly into phases and that's the point. The best 0 to 1 design happens when you stay close enough to the problem to let it keep surprising you.
The work doesn't separate cleanly into phases and that's the point. The best 0 to 1 design happens when you stay close enough to the problem to let it keep surprising you.
