Privacy

ONMIQ is a personal assistant, so the question is not whether it touches personal data — it is what it is allowed to do with it. This page is the short version of the model we actually implemented; the engineering detail lives in the project's PRIVACY.md and AGENT_SECURITY.md.

What ONMIQ stores

• Your account: email address and sign-in identity.
• What you told it to remember, and only that. Memories are suggested, never saved without your approval.
• Documents you upload, and the text extracted from them.
• Your conversations, tasks and agent plans, so you can go back to them.
• A content-free record of what the agent did: which tool, which operation, when, and whether you confirmed it.

What ONMIQ does not store

• Your Google password. ONMIQ never sees it; access is an OAuth token you can revoke.
• Full email bodies or document text in its activity log.
• Any credential in plain text. Provider tokens are encrypted and bound to your account.
• Sensitive fields in prompts by default: passport, national insurance, bank details and the like are never sent to the AI without a one-time approval for that single request.

Permissions

Reading your email does not give ONMIQ permission to send email. Each capability is its own switch, off until you turn it on, and every consequential action — send, change, delete — shows you the exact payload and waits for your confirmation.

The AI provider

Requests are processed by Anthropic's API. ONMIQ sends only the context a specific request needs, never your whole profile or mailbox.

Your controls

• Export everything as a file, from inside the app.
• Delete your account: every row tied to it is removed, not flagged.
• Disconnect Google: the token is revoked with Google where possible, and deleted locally either way — and we tell you which of the two happened.
• Turn any permission off at any time; a plan that needed it stops and says so.

Contact

Questions or a deletion request: hello@onmiq.app. We answer in English and Turkish.