Security

Local first, secure and reliable AI.

We believe in a local first principle. Using remote Embedding Models, Large Language Models and Whisper Models can be fast, cost efficient and depending on your use case a valid option. However by doing so you entrust these companies with your Data. You also pay per token. If you have large volumes of Data this can get very costly (see Cost) very quickly.

By going local first (or only), we can not only use the existing hardware for all of this for nearly free, but we also keep our data in house. By using our Custom Model Hosting feature you can also opt into hosting bigger models in a secure and central way for your users!