Models and limits
The connection is included
You do not need a provider account, an API key, or a card with an AI company. Your instance comes connected, and the cost is part of what you pay here.
Which models
A small set, because two kinds of job come up: a fast one for everyday questions, workflows and routine runs, and one that also reads pictures, scans and recordings, so a photographed invoice or a meeting recording is something you can ask about. Both are listed in your workspace wherever a model is chosen.
Today that is DeepSeek v4 Flash for everyday work and routine runs, and Gemini 2.5 Flash Lite for anything with a picture or a recording in it. Your workspace always lists the current pair wherever a model is chosen.
If you need a particular model
The model connection comes with your subscription, so there is nothing to set up, no keys to rotate and no second bill to reconcile. The list of models is ours to keep current, so you do not have to track what changed. If your team needs a particular model, or would rather use its own provider account, ask: it is a thing we can arrange.
The monthly ceiling
A workspace can be given a monthly ceiling on what it spends on models, so an agent stuck in a loop cannot quietly run up a bill. It is off until an administrator sets one in Settings, and the same screen shows how much of it has gone so far this month. An administrator can lower it whenever you want a tighter one, or ask us to raise it, and chooses what reaching it should do: refuse new runs until the month turns, or let the work carry on and send a warning instead. Nothing is lost either way, and work already finished is unaffected.