Short answer: very little. By design.
What we can see
- Your account information: name, email, subscription tier
- Your billing history: payments, invoices, plan changes
- Agent health signals: whether your agent is online, offline, or experiencing errors (anonymous technical status, not content)
- Usage metadata: the model name, the serving provider, the token counts, the cost and the time of each call, your plan tier, and technical fingerprints of the call, recorded by our model proxy and kept for up to 7 days. Never the content of those calls
What we cannot see
- Your conversations: messages between you and your agent are stored on your computer, not our servers
- Your files: documents, emails, calendar data, anything your agent touches lives locally
- Your agent's memory: all context and preferences are stored on your device
- The content of your prompts: prompts travel from your computer through Maxiful's model proxy to the AI provider. The proxy authenticates each request and enforces spend limits; it records only usage metadata (the model name, the serving provider, the token counts, the cost, the time, your plan tier and technical fingerprints of each call) and does not store the content of your prompts or the AI's responses
What about support access?
If you contact support and we need to diagnose a problem on your computer, we will ask for your explicit permission before connecting. You can approve, limit, or deny this at any time. Support sessions are time-limited and logged.
Our internal policy
No Maxiful employee accesses customer systems without explicit, documented consent. This is policy, and it applies to everyone on the team including founders.