Sovereign personal finance: transitioning to local-first budgeting.

Local-first budgeting means the ledger lives on your machine and answers to you. Not synced first and stored later, not held in trust by a company with a subscription plan: local first, and local only if you like. This guide covers what you actually gain, what you give up, how to carry your history across, and how to keep the method (envelope budgeting) intact while dropping the surveillance and the renewal date.

What "local-first" means when money is involved

In a cloud budgeting app the server holds the truth and your screen shows a copy. In a local-first one the file holds the truth and there is no server to disagree with it. That single inversion decides everything downstream: who can lock you out (nobody), who can read your spending (nobody), what happens when the company is acquired or shut down (nothing), and what you are paying for each year (nothing).

Self-hosted tools such as Actual Budget get most of this right and deserve their reputation. The remaining difference is the host: Actual expects you to run a server, or to pay someone to run one for you. Our version expects you to open a file. Both are local-first; only one of them asks you to become an administrator. The full comparison is here.

Envelope budgeting survives the move intact

The method is not the software. Envelope budgeting is four moves: record what came in, assign every unit of it to a named purpose, spend only from those names, roll the remainder forward at month end. YNAB taught a generation this discipline and then attached it to a bank connection and an annual bill. The discipline travels; the plumbing does not have to.

The Household Purse does the same four moves in one HTML file: envelopes, rollover, savings goals, a plain ledger, and a tax estimator. No feed, no categories guessed by a machine, no notifications.

The transition, in the order that actually works

1. Export first, cancel second. Pull a full CSV of transactions out of the incumbent while you still have an account. Mint taught everyone this lesson the hard way.
2. Take a month, not a year, of history. Most people import five years and use none of it. One month of real spending is enough to set opening envelope amounts honestly.
3. Write the fixed bills down once. Rent, utilities, insurance, subscriptions. Ten minutes here removes the only genuine convenience a bank feed offered.
4. Keep both running for one month. Overlap once, compare the two closing figures, then cancel.
5. Decide where the file lives. A synced folder if two people share it, a plain folder plus a backup if not. It is a file: treat it the way you treat a wedding photo.

What you give up, said plainly

Automatic transaction import. Push notifications. Investment dashboards that pull live balances. A phone app in a store. If any one of those is the reason you budget at all, stay where you are, honestly.

What you get instead is the small friction of typing a number, which is the part that changes behaviour, and a ledger that cannot be discontinued. See offline budgeting for the shorter version of this argument, and do the math for what five years of subscriptions actually costs.

Where sovereignty stops being an abstraction

Financial records are the most complete portrait of a life that exists in ordinary software: where you were, who you pay, what you are treated for, what you are planning. The argument for keeping that portrait off other people's servers is set out at length in Sovereignty as a Service and The Smaller Kingdom.

Every knife carries a thirty-day, one-e-mail refund, and the file stays yours either way. We could not reach into your computer to take it back if we wanted to. That is rather the point.

What to reach for