A companion you can actually trust with what you tell it.

A personal AI is only useful if you can trust it. You tell it the names of people you care about, the dates that matter to you, the things you're working through. If any of that leaks, gets sold, or trains someone else's model, the whole product is broken — not just for you, but for everyone who trusted it.

Syntrofos is built privacy-first on purpose. Not because privacy is a marketing checkbox, but because a companion that betrays what you tell it isn't a companion at all.

What privacy-first means here

Five concrete commitments shape how Syntrofos is built:

1. No ad tracking. Syntrofos does not contain ad networks, tracking pixels, or third-party analytics that build a profile of you for advertising. You will not see ads in the app.
2. No data resale. Syntrofos does not sell, rent, or share your conversations, memory, or personality settings with third parties. Not in aggregate, not anonymised, not ever.
3. No training on your conversations. Your conversations are not used to train foundation models for other products. Syntrofos is a product you use; the things you say to it stay with the product.
4. Full user control. You can see everything Syntrofos remembers about you, edit it, forget parts of it, or wipe all of it with a single tap.
5. No retention tax on leaving. Deleting your account removes your data within seven business days. There is no cooling-off period, no "we'll keep some of it for analytics," no dark pattern that tries to keep you.

These are commitments, not just policies. The privacy policy says them in legal language. The build says them in code.

How Syntrofos stores your data

A quick, honest summary of where things live:

• Conversations. Stored encrypted in transit and at rest, associated with your account. Used only to render your chat history and to inform long-term memory.
• Long-term memory. Stored separately from raw conversations. You can browse it, edit it, or wipe it without touching conversation history (and vice versa).
• Personality settings. Stored locally on your device and synced to your account so they survive reinstalls.
• Account email. Used for sign-in and important account notifications. Not used for marketing unless you opt in.

We don't keep anything we can't justify. If a piece of data isn't doing work for you, it shouldn't be sitting on a server.

What you can see and change

The privacy controls in Syntrofos are not buried. They're a top-level section of the app, and they're honest about what each control does:

• Memory browser. Everything Syntrofos has remembered about you, in plain language. Tap to edit. Tap to forget. Tap to wipe all.
• Conversation history. Every conversation you've had, with the option to delete individual ones or all of them.
• Personality data. The traits, tone, and humour settings you've chosen. Edit anytime; changes apply on the next conversation.
• Account. Email, sign-in method, and the option to delete the account entirely.
• Export. Take your data out — conversations, memory, personality settings — in a single export, in a format you can read.

This last one matters. Privacy isn't real if you can't leave.

How deletion works

When you delete your account, here's what happens:

• All conversations are deleted from production storage within seven business days.
• Long-term memory is deleted within the same window.
• Personality settings are deleted from your device and the sync layer.
• Backups rotate out within thirty days, per the standard backup retention.
• Confirmation email is sent when the deletion is complete.

If you delete Syntrofos from your device without deleting your account, the data is still on the server. If you want to wipe the server side, use the in-app account deletion. There's also a web form on the Delete Account page for situations where you can't open the app.

What's not in scope

A few things Syntrofos deliberately doesn't do, because doing them would compromise the privacy model:

• No cross-device sync of raw conversations. Conversations are tied to the device where they happened, plus the server. They don't sync peer-to-peer or get backed up to third-party clouds you didn't choose.
• No social features. Syntrofos has no friends list, no "share this conversation," no leaderboard. There's no social layer to leak into.
• No voice training on your voice. If voice features are added later, they will be opt-in and the audio data will not be retained for model training.
• No third-party AI in the middle. Syntrofos runs on a backend we control. There is no "powered by" partner whose privacy policy you also have to read.

These omissions are the privacy model. Without them, "privacy first" would be a slogan.

Why this matters for a personal AI

A personal AI sits closer to the user's life than almost any other category of software. The things people say to a companion are the things they wouldn't say to a search engine. If that trust is broken — even once — the category is hurt for everyone.

Syntrofos is built around keeping that trust. Privacy isn't a feature you can toggle. It's the substrate everything else depends on.