What a cloud app learns from a focused person
Local-first software is software that stores your data on your own device instead of a company's server, and that one architectural choice decides who your focus data really belongs to. When your productivity app lives on someone's server, the operator can see more than your task list.
They can see when you start work and when you stop. How long your focus blocks really last. Which goals you set and quietly abandon. What you type into a journal at the end of a hard day. Individually these are small things; together they are a behavioral profile of your working life.
Most companies will never misuse this. That is not the point. The point is that the data exists, it accumulates, it outlives your subscription, and its future uses are decided by a privacy policy you have not read since it last changed. The EFF's privacy work documents how rarely 'we would never' survives contact with an acquisition or a new business model.
It is not malice, it is incentives
Nobody at a productivity company wakes up wanting to surveil you. The attention cost of these systems is real, as the productivity-app research shows. But every cloud tool is run by a business, and businesses optimize what they can measure. Engagement curves, retention cohorts, feature adoption: all of it is computed from your behavior. Even a well-meaning team ends up asking: what makes people come back? And the honest answers to that question are rarely calm.
This is how you end up with streaks that make you anxious, notifications timed to pull you back, and dashboards that turn your attention into someone else's chart. Not because anyone is evil, but because the architecture makes it easy and the incentives make it likely.
Notice, too, that none of these features could exist without the data leaving your device. A streak requires someone else to count. A 'your team is ahead of you' notification requires someone else to compare. The engagement machinery is not an add-on to the cloud model; it is the cloud model's most natural expression. Remove the server, and the temptation is gone by construction rather than by policy.
What local-first means, concretely
Local-first is a simple architectural promise, laid out in the Ink & Switch essay that named it: your data is stored on your device, and it is not sent to a company's server by default. In Focus, that means the one-intention-per-day practice, its focus blocks, its stats, and its evening reflections live on your phone. There is no account holding them, no analytics pipeline reading them, no dashboard anywhere else with your name on it.
Local-first means your data is stored on your own device and never sent to a company's server by default. If the company disappears tomorrow, your data does not notice. If you lose signal, nothing about your day stops working. The app is a tool you hold, not a service you are rented.
The honest tradeoffs
I would rather tell you what this costs than pretend it is free. Local-first means no magical sync across five devices; sync is the single hardest problem in this architecture, and anyone who says otherwise is selling something. It means if you lose your phone without an export, your history goes with it. And it means the company (me) cannot fix your data from a support console, because I cannot see it.
These are real costs. I accept them because of what they buy: the inability to betray you. A company that does not hold your data cannot leak it, sell it, train on it, or be compelled to hand it over. Constraints are the only promises that keep themselves.
It is worth saying plainly who local-first is not for, at least today. If your work lives across three devices and a shared family iPad, a cloud tool will simply fit your life better right now. If you want your team inside the same workspace, local-first single-player software is the wrong shape. Choosing a tool is choosing which problems you would rather have; I am only arguing that most people have never been offered the other column.
A note on our own waitlist
In the spirit of not asking for anything I would not give: this website does collect one piece of data, your email address, if you join the waitlist. It is used for access and rare product notes, nothing else, and you can have it deleted by replying to any message. That is the entire data practice of the site, and it is written down in the privacy policy in plain language.
The line I draw is this: the website may hold your address because you handed it over to be contacted. The app holds nothing, because your attention was never mine to keep. Where the line sits is a judgment call every tool maker makes; what matters is that it is drawn somewhere visible, and you are told where.
The export test
Whatever tool you use, apply the export test: can you leave, right now, with everything, in a format you can actually read? If the answer is a JSON dump behind a support request, that is a rental wearing a purchase's clothes. Your record of your own attention should be yours in the most boring sense: files, on your device, openable in any text editor.
That is the whole argument, really. Your focus data is the diary of your working life. Diaries belong in drawers you own, not in buildings you visit on permission. Focus keeps it there: on your device, exportable whenever you ask.
- Local-first means your data is stored on your own device and never sent to a company's server by default.
- A cloud productivity app can observe when you work, how long, what you abandon, and what you avoid: a behavioral profile, not just a task list.
- The tradeoffs are real: local-first costs you effortless sync and cross-device access.
- The export test: a tool that respects you lets you leave with everything, in a format you can read.