Questions, answered plainly

Because 30% is a guess that happens to be roughly right for a middle slice of people and meaningfully wrong for everyone else. It ignores your filing status, your standard deduction, the qualified business income deduction, whether your state taxes income at all, and whether you have a salary that already has tax withheld from it.

Set aside too much and you’ve loaned the government money interest-free. Set aside too little and you find out in April. We’d rather calculate it.

It’s Social Security and Medicare. When you have a job, you pay half and your employer pays the other half. When you work for yourself, you’re both, so you pay both halves, 15.3% on most of your net earnings. It’s separate from income tax and it’s usually the biggest surprise for people in their first year of freelancing.

The IRS expects tax to be paid as you earn it. An employer handles that by withholding from each paycheck. Nobody does that for you, so you send payments yourself four times a year. In 2026 those dates are April 15, June 15, September 15, and January 15, 2027. Miss one and interest starts accruing on that quarter.

The qualified business income deduction lets most sole proprietors deduct 20% of their business income before income tax is calculated. It’s a large, real reduction that flat-percentage tools can’t account for, because they never look at your income picture in the first place. We apply it, and at higher incomes we reduce it gradually, the way the law phases it down.

A rule that protects you from an underpayment penalty. If your payments across the year reach 90% of what you end up owing, or 100% of what you owed last year (110% if last year’s income was over $150,000), you won’t be penalised, even if you end up owing more at filing. It’s especially useful in a year your income jumps: you can pay against last year’s smaller number and settle the difference in April.

Pay the missed amount as soon as you can rather than adding it to your next payment. The IRS charges interest on each quarter’s shortfall from that quarter’s own due date, so waiting only lets it grow. When a quarter is underpaid, your dashboard shows the catch-up amount to pay now, separately from the next deadline.

No, and that’s deliberate. This product is read-only. It cannot receive, hold, or transfer your money. Not as a disabled feature, but as something that was never built.

We tell you how much to set aside before each quarterly deadline. You move it, and you pay the IRS yourself, so your money never leaves your own accounts.

Moving money means licensing, custody obligations, and a much larger surface for things to go wrong. We wanted to prove the calculation is worth having first.

No. Plaid handles authentication directly with your bank. Your credentials never touch our servers. We receive a token that can read transactions and nothing else, and that token is encrypted before it’s stored.

Rougher than the federal one, and we say so wherever we show it. We model each state with a single rate rather than its full bracket structure. For the flat-tax states that’s exact; for progressive states like California and New York it uses the top marginal rate, which overstates tax at typical freelance income levels. Nine states have no income tax at all, and those come out at zero.

Yes, the ones you categorise. Spending from your connected accounts that you mark as a business expense, like software, equipment, or travel, is deducted from your income before the estimate is calculated, and business meals count at 50%, as the IRS allows. Anything you haven’t categorised isn’t counted, so the estimate errs on the high side until you do.

Today, you do, or your own accountant does. Setaside doesn’t file returns. We’re working toward a partner CPA who files your return straight from your Setaside records, for one price for the tax year. That isn’t built yet.

No. It’s an estimate to help you plan, not advice and not a filing. For anything specific to your situation, talk to a tax professional. This tool replaces guessing, not them.