Privacy
Resident is built so that what you say to it stays on your device. This page says exactly what that covers.
What stays on your device
- Your chats, including attached photos and documents.
- The models you download, and everything they compute.
- Memory: the facts the agent has saved about you. You can read, edit or clear them in Settings, or switch memory off.
- Skills and agents you create.
- What you dictate. Speech is turned into text by a model on the device.
Resident has no account, no analytics and no server of its own. We cannot see your chats, because they are never sent to us.
What leaves your device, and when
Only what a tool needs, and only when the agent uses that tool:
- Web search sends the search terms to a search service.
- Reading a page requests that page’s address from the site that hosts it.
- A connector (Linear, Notion and others) is sent what the tool call asks for, to the service you connected. Signing in to a connector happens with that service directly; Resident stores the resulting access token in the device’s Keychain.
- Downloading a model fetches its files from Hugging Face.
- A server you add, such as Ollama on your Mac, receives the chat you have with it.
Each of the first three has a switch in every agent. With them off, nothing you write leaves the device.
Apple Intelligence
If you choose Apple Intelligence as the model, your request is handled by the model built into your device by Apple, under Apple’s own terms.
Changes
If this page changes in a way that matters, the change will be described here with its date.
Last updated 16 November 2026.