An absence, not a setting
On Android, owed is being built without the internet permission. On iOS, there is no networking code. The details differ by platform, but the decision is the same: the ledger has no network route out of the app. This is not a switch a person has to remember to turn off. The capability is absent from the design.
That is a specific privacy claim, and it is stronger for staying specific. A local app cannot answer every question about privacy. Someone with access to an unlocked phone can still see what is on it. A file you choose to share can still be seen by its recipient. But the app itself cannot upload the ledger over the internet. We think a boundary is more useful when it can be described in plain terms and checked against the software.
Why we care this much
A list of informal debts can look minor beside a bank statement. It still records personal things: who was involved, what was owed, whether it had been settled, and when a follow-up might be appropriate. The amounts may be ordinary. The relationships around them are not just fields in a form.
Most people do not need a server to make a short ledger useful. They need a place to put down the detail before it slips from memory, then a clear view of what is still open. If those records stay on the device, a service does not need to receive them just to display the board. The notes remain available on the device where they were made.
Cloud services can make information easy to reach across devices. That convenience is real. It also means copies must be stored somewhere, access must be managed, and the service becomes part of the trust decision. For this particular job, we kept asking what the connection would add. We could not find a reason strong enough to justify sending the ledger anywhere.
A small ledger, with clear edges
The board is for keeping track of who owes what as an item moves through sent, due, and paid. Those states give the list a simple shape. Instead of trying to remember each conversation in full, a person can look at the board and see what needs attention. The point is not to turn a personal exchange into bookkeeping. It is to make the practical detail easier to hold.
When a follow-up is due, owed provides polite draft wording. A draft is a starting point, not a message sent on someone's behalf. You can change the words to fit the relationship, decide whether to use them, or leave them alone. That distinction matters: a tool can reduce the effort of finding a gentle sentence without deciding when a conversation should happen.
Reports offer another way to review the records. A JSON backup gives you a file you own. You can keep it where you choose and carry a copy outside the app. That makes the record less dependent on a particular service, while also putting the ordinary responsibility of backups in your hands: a file only helps if you make it and keep it somewhere safe.
The trade-off we chose
There is a cost to keeping the ledger local. There is no cloud sync, so a second device will not receive an automatic copy or stay in step with the first. If you move to a new phone, the backup is a deliberate part of taking your records with you. If you never make one, there is no remote copy to restore. We would rather name that limit than promise the ease of sync without acknowledging what it requires.
This choice does not make local storage universally better. Different software has different jobs. A shared calendar, for example, may need to coordinate several people; a personal record may not. We are making small on-device apps, and we want their connections to follow from a clear need. For owed, the useful work happens on the device. The network does not improve that work enough to earn a place.
owed has not launched yet. We are still building it around that boundary: a local board, useful ways to review it, and a backup you control. The current outline is on owed.