The rapid growth of AI, and the data centres needed to support it, is a real and important issue. It can affect energy grids and local water supplies, and we take that seriously.

That said, we believe AI has a genuine role to play in making climate and sustainability information more accessible. The challenge is doing it responsibly.

Here is how we try to get that balance right:

We use Infomaniak as our AI provider. Infomaniak is a Swiss-based company widely recognised as one of the most ethical cloud and AI providers in the world. They run their infrastructure on renewable energy and are transparent about their environmental commitments.

We use lightweight, efficient models. Not all AI models are equal when it comes to resource use. The most energy-intensive part of AI is training large models from scratch, not running them day to day. We use Mistral and Apertvs, two models specifically built to be compact and efficient.

Apertvs is fully open-source, developed by ETH Zurich and EPFL, two of Europe's leading research universities. It is released under an open licence, which means its training, architecture, and data are publicly documented. No black box.

Your data stays in a controlled environment. Because we use Infomaniak's infrastructure, your queries are processed in a closed system, not sent to large commercial cloud platforms.

We are not perfect, and the field is evolving fast. But we believe choosing the right partners and tools makes a real difference.

The AI Companion draws on two sources, and it always tells you which one it is using.

First: our own curated knowledge base. We have built and maintain a wiki of climate and sustainability information, sourced, reviewed, and validated by us. When you ask a question, the Companion first searches this wiki using MistralAI. If a clear answer is found there, that is what you get.

Second: a fallback to broader AI knowledge. If the answer cannot be found in our wiki, the Companion falls back to Apertvs, an open-source model developed by Swiss research institutions ETH Zurich and EPFL, to provide a response based on its training data. When this happens, the answer is clearly labelled so you know it is coming from general AI knowledge rather than our curated sources.

This approach means that, as much as possible, answers are grounded in verified, local information, and you always know where they come from.

This app is not trying to capture your attention or keep you scrolling. We are aware of the pressure that comes from constant digital noise, and from the weight of negative climate narratives. That is not what we want to add to your day.

Our research shows that people want simple, trustworthy ways to stay informed, and that short, playful challenges are a good way to do that without feeling overwhelmed.

With that in mind, Environmentle is built around very short, casual challenges. You can take 2 to 3 challenges a day, for a maximum of 3 to 5 minutes. That is intentional. We want you to get something useful out of it, and then get on with your day.