Principles
- Each agent runs separately — one agent’s data isn’t available to another
- Passwords and keys are stored encrypted — nobody but you has access to them
- You decide what the agent can do — sending emails, deleting files, creating events: each action requires your permission
- Access can be revoked at any time — disconnect any service or delete the agent completely
What the agent doesn’t do without your permission
- Doesn’t send emails
- Doesn’t share your data with other users
- Doesn’t use your data to train AI models
Where to manage this
Permissions
Choose what the agent can do on its own and what only with your confirmation.
Secrets and keys
Secure storage for service tokens and keys.
Danger zone
Reset the agent and permanently delete all its data.
Memory
What the agent remembers about you and how to delete it.

