We are now open!
You can now start using our open APIs to use language models with support for European languages. All without relying on foreign infrastructure.
We are very happy to receive feedback on how we can improve — we want to build the service together with our customers!
The models we host are all be so-called open models. We aim to have a mix of a few reasoning models, and some smaller models that are good for handling, for example, RAG pipelines and answering simpler questions. In addition, we will have an embedding model and a reranking model. We will also be quick to offer new open models as they are released - and aim to support European and local models to unlock AI innovation in multilingual contexts. Those that prove to be popular with our users will remain, while those that are less popular will be replaced with newer, better ones.
No. All data sent to our API stays in Sweden. We also don’t store any data on our side except what is absolutely necessary, such as logs for troubleshooting or to prevent abuse. Our GPU servers are not allowed to communicate with the Internet, so you can be completely confident using our services for sensitive data. Additionally, all traffic is of course encrypted. Please read more about how we ensure high security in our privacy policy.
Yes and since we know that some companies and public sector organizations prefer to pay by invoice, we offer a prepayment setup where you purchase so-called tokens in advance via invoice. We support sending invoices via Peppol. This makes it easy and convenient to keep track of your costs.