I wanted to see my own AI history.

First, a place to see my own AI history.

AI is amazing. I was watching people use it constantly and create huge amounts of software. I wanted to understand and compare what they were building, but I couldn’t find a clear record to look back through.

I wanted all my AI usage in one place so I could track it, connect it to the work I was doing, and revisit past events. Which tools did I use? How much did I use them? What was I working on?

Then, the footprint behind the work.

I started researching AI’s environmental impact and wanted to know how much energy my own usage consumed.

I spent a lot of time digging into that question. I wanted my own numbers, but I was still figuring out how to track them and what the tool would look like.

Then, the name gave the idea shape.

I started connecting two ideas: a central place for my AI, and a contrail. Bringing central AI and contrail together gave me the name Centrail.

Thinking about that connection led to a realization: AI work leaves a trace. An aircraft leaves a contrail; AI activity could leave a record of usage, consumption, and the work associated with it. That was the “AI contrail” I wanted to be able to follow.

I couldn’t find that kind of record, so I set out to build it. What began as a search for my own numbers became a way to peer into the history of my AI usage. We call that trace a centrail.

That became GreenAI’s starting point. Energy, water, and carbon are modeled estimates. Read the methodology here.

A personal question became a team problem.

While working in corporate environments, I kept hearing the same frustration: teams couldn’t clearly monitor their AI costs or understand how to bring them down. The record I was building felt like a starting point for that problem, too.

The data could help teams look back at their usage, see where they were spending money, and connect it to repositories and commits. The tool I was already building gave me a place to start.

Today, Centrail tracks usage from supported coding-agent logs and connects tokens and estimated cost to Git history.

From AI contrails to agent economics.

The more I built, the more excited I became about what this could do. I began thinking about agent economies and how much we would need a reliable record of AI work. The idea grew into building a real source of truth and a ledger for AI work.

What started with tracking my own usage had become something I wanted to help other people use. I wanted people and teams to be able to see their AI history, understand its costs, and make decisions with data they could actually look at.

We’re proud of what we’re building with Centrail. We want to help change the AI world and make a difference where we can, from how teams spend their budgets to how they understand the resources their AI uses. There’s a lot ahead of us, and I’m excited to keep building it.