Web Experts builds tools powered by AI. That comes with a responsibility to understand the environmental impact of the technology we are using and putting out into the world, not just the parts that are convenient to think about.
AI has a carbon footprint. The more we looked into it, the more we realized how difficult it was to figure out what ours actually was.
Why Existing Calculators Fell Short
There are some very good AI environmental calculators already available. We are not knocking them, and we spent real time with them before deciding to build anything.
The problem is that most of them measure only pieces of the footprint. You can get a number, but the number answers a narrower question than the one we were asking, and stitching several partial answers together is not a great way to understand your own impact.
What TokenPrint Actually Does
TokenPrint estimates energy, carbon, and water. It accounts for model training and hardware manufacturing, which are the parts that tend to get left out of simpler tools, and it considers where your electricity comes from, because that changes the answer considerably.
It also helps you figure out how much AI you are actually using in the first place. That turns out to be the hardest input for most people to supply, and a calculator that assumes you already know is not much of a calculator.
Free, No Signup, Methodology Published
TokenPrint is free. It requires no signup and no email address, because asking for contact details in exchange for an environmental estimate felt like the wrong trade.
All of the methodology and sources are published. If you disagree with an assumption we made, you can go look at it and argue with it, which is how a tool like this should work.
What Building It Taught Us
Building TokenPrint forced us to look more closely at how we use AI in our own development work. That starts with the models we choose, but it does not stop there.
It extends to the coding tools our developers use every day, and to the AI features quietly built into software we already had running. A lot of consumption is invisible until you go looking for it, and we found plenty we had not accounted for.
Know It Before You Reduce It
Before we can reduce our AI footprint, we need to know our TokenPrint. Reduction targets set against a number nobody has actually measured are just a feeling with a percentage attached.
This is what ethical web development in the age of AI looks like from where we sit. Measure first, then decide what to change.