In August 2025, I was laid off.
That same day, I made a decision: I didn't want to go back to the traditional corporate world. After years of reorganizations, management changes, shifting priorities, and the constant possibility of being laid off again, I wanted something different.
I wanted to build my own product — something I could own end-to-end, be accountable for, and shape according to my own vision.
Like many people, I turned my attention to the stock market. I had a theory I wanted to explore and believed I could build a system around it. Easy money, right?
To get started, I sat down with one of my closest collaborators at the time: ChatGPT.
In the beginning, the workflow was simple. GPT would generate code in chat, I would copy it into Visual Studio Code, run it, report the results, and we'd repeat the process. One iteration after another.
Eventually, I discovered that AI agents could run directly inside Visual Studio. Suddenly, I no longer had to copy and paste code between windows. I could interact with the agent directly in my development environment and focus on guiding the work instead of moving text around.
It felt like a major leap forward.
Then the real challenge appeared.
As my project grew, so did the number of ideas, experiments, and directions I wanted to explore. AI made building dramatically faster, but it also created a new problem: managing everything that was happening.
Around this time, Claude was becoming increasingly popular, so I decided to see what all the excitement was about. I installed Claude CLI and immediately saw the appeal. For the first time, I could run multiple conversations in parallel, each focused on a different part of the project.
That worked well — until it didn't.
Soon I found myself with multiple terminal windows open, each containing a different conversation. I'd finish for the day, leave everything running, and return the next morning trying to remember which agent was working on what.
Context became a problem.
Some agents had no memory of previous work and needed to re-analyze project files to understand where things stood. Documentation was inconsistent. Notes were scattered. New ideas were arriving faster than I could organize them.
More often than not, I found myself explaining the same things repeatedly — what had already been built, why certain decisions were made, and what needed to happen next. If I forgot to ask an agent to document its work before ending a session, much of that context was effectively lost.
The result was chaos.
My original stock market project had already expanded into multiple related projects, each with its own goals, experiments, and requirements. I was juggling conversations, notes, documents, tasks, and ideas across all of them.
And then it hit me.
I was spending more time managing agents than building.
That was the moment I realized the real problem wasn't coding.
The real problem was coordination.
I searched for existing tools that could solve this problem, but I couldn't find anything that matched the way I wanted to work. Most solutions focused on individual conversations or individual agents. What I wanted was a way to manage the entire ecosystem around them.
So I decided to build it myself.
That project eventually became Clayrune.
Originally, it was called "Mission Control" — a name suggested by Claude. The goal was simple: create a central place where I could see everything happening across my projects.
I wanted a visual overview of active work, the ability to run multiple AI conversations in parallel, a backlog where I could store ideas for later, and a clear understanding of which agent was doing what at any given moment.
As development continued, Clayrune gradually became more than a tool for managing my investment project. It became a project of its own.
Features and ideas continued to pour in, but now they had structure. Work became organized, documented, and manageable. I could maintain separate projects with their own rules, instructions, and operating procedures. I could schedule recurring tasks, automate workflows, monitor progress, and let agents handle routine work independently.
Most importantly, I could finally focus on building instead of managing chaos.
Because Clayrune runs on your own machine and uses your own agents, I could continue using my existing subscriptions and workflows without being locked into a specific platform or worrying about additional usage costs.
From the very beginning, I knew Clayrune would be open source.
The goal was never to build a closed platform. The goal was to create something useful and share it with others who face the same challenges I did. If Clayrune helps you stay organized, build faster, or manage your AI workflows more effectively, then it has accomplished exactly what it was designed to do.
Development continues every day, and there is still plenty left to build.
AI agents are becoming more capable every month. The challenge is no longer getting them to do work — it's keeping that work organized, visible, and moving in the right direction.
Clayrune is my answer to that problem.
I hope it helps you build something great.
— Ron