Security & Privacy

Your chats stay yours.

Freffo reads WhatsApp so you don't have to scroll: here is exactly how that stays private, in plain language, with nothing overclaimed.

One isolated instance per person

Your data lives in its own private instance with its own storage and its own WhatsApp session, stored in India. It is never pooled with other users, never sold, never shared.

WhatsApp's encryption stays intact

Freffo connects as a linked device, exactly like WhatsApp Web. Nothing about WhatsApp's own end-to-end encryption changes, and you can unlink it from WhatsApp at any moment.

Every chat has a switch

You choose what the assistant may read. Switch a chat off, from the dashboard or by just telling Freffo, and it is never sent to the AI again, and what was already pulled from it is deleted a day later.

Never used to train AI

Your messages are processed to serve you, and that is all. Nothing you or your contacts write is ever used to train AI models, ours or anyone else's.

Nothing sent without your OK

Freffo drafts the message and shows you the exact text: it goes out only when you say "send", or at the time you set, and only to the person or group you named. Everything else it says stays in your own Freffo chat.

Leave anytime, data deleted

Unlink Freffo in WhatsApp → Linked Devices (it shows there as Google Chrome or Firefox) and it stops at once. Ask us to delete your account and its messages, media and connections are deleted (only a short record of it is kept, as the Privacy Policy says), and Freffo leaves your Linked devices. A copy kept in case of a mistake, without your media, is deleted after 30 days.

What the AI actually reads, honestly

To build your briefs and reminders, Freffo reads the chats you leave switched on and uses AI to summarise them for you. That is the only reason it looks at them. Chats are on by default so the assistant is useful from day one: the switches above put you in control of exactly what it sees.

Third-party messages are treated as data, never as instructions: something a contact writes cannot command your assistant, message anyone, or change your settings. Each instance authenticates with its own per-user credentials, sign-in is brute-force capped, and the assistant that answers in shared chats runs with read-only access and no view of your private memory.