Shared knowledge that stays yours
The question is rarely "how do we share notes". It is what happens to internal knowledge once it is somewhere you do not control.
The server never sees plaintext
A team vault is encrypted before it leaves the machine, and sync moves ciphertext. Whoever hosts the storage — including us, if you use ours — holds bytes they cannot read.
Membership is a key, not a checkbox
Adding someone grants their key access; removing someone re-encrypts what they can no longer open. Access is enforced by cryptography rather than by a permission field somebody could flip.
Your knowledge is not training data
Nothing is uploaded to us for analysis and there is no model being improved on your content. If you connect a cloud model, that provider's terms apply to what you send it — which is why pointing it at a local model is a real option here and not a slide.
Legal asks what happens to the handbook
It contains salary bands, an incident write-up, and two customer names.
The vault is encrypted at rest and in transit, and the sync target holds ciphertext only. Roles say who can write and who can read. The audit of who wrote what is signed. If the answer required is that no third party can process the contents at all, the AI runs against a local model and nothing leaves the building.
The conversation stops being about trust and becomes about architecture, which is a conversation you can finish.
Can I use this with my data?
Retrieval is local. Generation goes where you point it.
Finding the right passages happens on your machine, always. What happens next is your choice: connect a cloud model and the retrieved passages travel to that provider, exactly as they would if you pasted them; point it at a model running locally and nothing leaves the machine at all. That second option is available today, and it is what turns "where is our data processed?" into "on your device".