Environmental impact
Software that calls a model uses energy. We would rather say plainly what we do about it than make a claim we cannot stand behind.
Why it matters to us
Every time Maestra calls a model, that request runs on a computer somewhere and uses power. For a company whose product is built around those requests, that is a real cost to account for, and one worth thinking about from the start rather than later.
So we design the product to ask for compute only when it earns its keep, and to keep as much work as possible on the machine in front of you.
What we do about it
We use efficient models matched to each task, rather than the largest model available. A smaller model that does the job well uses less compute.
A request is sent only when you ask for one. Nothing runs in the background, and the app is not calling a model while you work unprompted.
Style analysis, preset building and audio measurement run on your own computer. That work does not travel to a server at all.
Maestra ships as lightweight desktop software you run when you want it, rather than an always-on cloud service that consumes resources around the clock.
The requests that do leave your machine are handled by our provider in the EU, which keeps nothing after the reply.
What we do not claim yet
We have not measured our footprint, so we are not going to put a figure on it. We make no claim to be carbon neutral, we hold no certification, and we do not buy offsets. We would rather leave this blank than fill it with a number we cannot support.
When we can measure it properly, we will publish what we find here, whatever it says.
As the product grows, we will keep choosing efficiency over scale wherever the choice is ours to make.