Skip to content
Docs
Enterprise

Cascadia Enterprise lets an organization run AI models across many of its PCs and manage them as one service. It builds on the open-source runtime and adds what a team needs to run it at scale:

  • Add devices without setup work. Each PC enrolls itself from a one-time installer, or from a link you send to whoever uses it.
  • Deploy models from one place. Choose a model and the devices to run it on, as whole copies or split across several PCs.
  • Control who has access. Issue API tokens for your applications, and revoke any token, device or member within about a minute.
  • Stay up to date without visiting machines. Devices install signed updates themselves and roll back if one fails.
  • Work as a team. Invite colleagues as members, admins or owners.

Models run on your machines, and your applications reach them through a single OpenAI-compatible API. Prompts and replies stay on your network.

Piece What it does Where it runs
Console Your account at console.cascadia.to allows you to manage members, devices, API tokens and billing. It also hosts your fleet’s certificate authority (CA), which decides which machines belong to your fleet. Hosted by Cascadia
Devices Machines running the Cascadia node service. Each one serves models and joins the others in your fleet. Your Windows and Ubuntu PCs
Fleet Manager A dashboard for watching and controlling the fleet from inside your network. Deploy models, restart devices, roll out updates. An admin’s computer
Gateway The OpenAI-compatible API on each device. Applications on that device call it with a token from the console, and the fleet routes each request to devices serving the model. Every device

You add devices from the console, and each one enrolls itself. There is nothing to approve by hand and nothing to receive from Cascadia.