Where suum-exse.ai gets its name
suum ex se
/ˈsu.um ɛks ˈseː/
soo‑oom ex say · Latin
Latin phrase · “one’s own, from oneself”
suum-exse is AI built on well-known open-source models. Your data is encrypted, never sold, and never used for anything — including training — unless you say yes. Download every last piece of it to your own machine whenever you like. We encourage it.
Who decides what happens to your data?
suum-exse
You decide. Always.
The default is no. Nothing is sold, and nothing is used beyond answering you unless you say yes. All of it downloads to your machine, any time, no questions.
“Opt-out” AI companies
They decide — unless you catch it.
Your data is used until you find the setting to stop it — and it stays off only as long as their policy says so.
Conventional AI companies
They decide.
Your data trains their models by default, under terms they can revise at any time.
Protection, in two levels
We would rather tell you exactly where we stand than promise what we cannot deliver yet. Here is how your data is protected at launch — and what every reservation helps us reach.
Level I · Live at launch
Everything you keep with us — your conversations, your memory vault, and every app we add, from email to calendar — is encrypted, never sold, never shared, and never trained on. The default is no. Download all of it, any time.
Level II · At $5,000,000
At launch, the AI runs on well-known open-source models, served through a trusted provider’s API. When you ask something, that request — and any part of your data it needs — goes to them to be answered. We send only what the question needs, but once it is there, their terms apply, not ours. At $5,000,000 we buy our own hardware and run those same models ourselves. From then on, nothing leaves our hands.
Why start this way? It lets us open sooner, at a fair price, on open models anyone can inspect — and every plan reserved moves all of us closer to Level II.
The waitlist Sample numbers
0
Pro yearly plans reserved
0
Max yearly plans reserved
$0
Reserved toward Level II — every model on our own hardware
Why $5,000,000: it is what it takes to buy our own hardware and run every model ourselves — Level II, where nothing leaves our hands. How the two levels work.
The idea
Your data is encrypted and we never sell it. Using it for anything beyond serving you — training included — takes your explicit permission, and you can take that permission back.
Download everything to your own machine, in open formats, whenever you want. No request form, no waiting period, no limits. Keep your own backups — we would rather you did.
Well-known open-source models anyone can inspect — served through a trusted provider at launch, on our own hardware at $5,000,000, and in time, models of our own.
What you get
Most of us keep our mail, calendars and notes with companies none of us ever really trusted. We are replacing that software one app at a time — each one built to be genuinely good, and each one answering to you.
A fast, capable assistant that remembers your work. This is what ships first.
Everything your AI remembers about you, encrypted and kept in one place — yours to open, edit, download or erase.
Your inbox, drafted and triaged by an assistant that never mines, sells or shares your mail.
Scheduling that understands the week you actually have, not the one on the grid.
A list that keeps itself, pulled from what you said you would do.
Kanban boards and Gantt timelines for your own work or your whole team — organised by project, with plans, documents and decisions held together with the thread of why.
Recording, transcripts and notes on every plan — kept encrypted and kept yours. Download them, delete them, or keep them forever.
“Suum cuique tribuere.”
To render to each what is theirs. — Cicero, De Officiis, I.15
Why the name
Cicero set this down as a duty of justice, and Roman jurists later put it at the head of their law. It assumes something worth defending: that what is yours should be returned to you, not kept by whoever happens to be holding it.
We hold your data — that is how the cloud works. But we hold it the way a safe-deposit box holds what is inside: locked, yours, and handed back whenever you ask. We took the name as a standard to be held to.
Read the missionQuestions
With us, in the cloud, encrypted in transit and at rest — and always downloadable to your own machine. That is Level I, in place from launch. The one thing that travels further for now is each question you ask the AI, which is answered through a trusted provider’s API until we reach Level II (see “Who else sees my data at launch?”).
Yes — everything, whenever you like, in open formats you can read without us. No request form, no waiting period, no limits on how often. We actively encourage keeping your own backups on your own machine.
Only if you explicitly allow it. The default is no. Permission is opt-in, specific, and yours to withdraw at any time.
We never sell it, rent it, broker it or hand it to partners. Your stored data is never shared — that is Level I, from day one. Your questions to the AI are answered through a trusted provider until we reach $5,000,000 and run the models on our own hardware — that is Level II (see the next question). Beyond that, if your data is ever shared with anyone, it is because you chose to share it.
At launch, the AI runs on well-known open-source models, served through a trusted provider’s API. When you ask something, the request — and any part of your data needed to answer it — goes to that provider to be answered. We send only what the question needs, and nothing you store is handed over beyond that — but once a request reaches the provider, their terms govern it, not ours. That is exactly why Level II matters to us. When reserved plans reach $5,000,000, we buy our own hardware and run the same models ourselves: Level II, where nothing leaves our hands. See the two levels and our progress.
Well-known open-source models, chosen by us and updated as better ones arrive. This is not a bring-your-own-model service. At launch they run through a trusted provider’s API; at $5,000,000 we run them on our own hardware; after that, we build models of our own. On our side, the rules never change: your data is used to answer you, and nothing else.
If you have been keeping backups — and we will keep nudging you to — you already have everything, in formats that work without us. That is exactly why we make downloading easy.
You choose Pro or Max, monthly or yearly, and pay your first month or year to hold your place. It does not start until your access opens, and you can get a full refund any time before then. We open access in small batches, starting with chat and the memory vault.
Not right now. Every account runs on models we pay to run, and we will not pay for free accounts with anyone’s data.
Some apps may carry ads. If you want ads targeted to you, you can opt in, app by app. The default is no.
Every purchase is a statement to the companies that have been poor stewards of our data.
Join the waitlistFrom $8.33 / month, billed yearly · refundable until access opens · never sold
