A non-technical founder can now ship a working SaaS in a weekend with Cursor, Claude Code, and a handful of API keys. Indie builders launch three products a month. Twitter is full of $0-to-$10k MRR screenshots from people who haven't hand-written a line of code in six months.
The AI is doing the engineering. What it isn't doing — and what almost nobody seems to be thinking about — is the legal structure underneath the product. You can vibe code the app. You cannot vibe code the liability.
The standard indie launch in 2026
Domain bought on a personal card. Stripe connected to a personal bank account. App deployed under a personal name. The product ships, the first dollars come in, and the founder moves on to the next build. It works — until it doesn't.
When it stops working, the consequences land on the founder personally, because there is no company to absorb them. The patterns are predictable:
- A user sues. The AI-generated app mishandled their data, charged them twice, or shipped a feature that infringes a patent. The complaint names you. House, savings, and personal accounts are all in scope.
- Stripe freezes the account. A chargeback spike, a compliance flag, an angry support email — and 90 days of revenue is locked. With no business entity, there's no leverage and no clean path to migrate elsewhere.
- The bank closes the account. Most retail banks explicitly forbid using personal accounts for business income. They detect it eventually. When they do, they close the account, and depending on volumes and jurisdiction they may also report the activity to their AML unit.
- The tax authority writes. Receiving SaaS revenue as an individual in most countries means income tax plus social charges on the full top line, with limited deductions. By the time the founder realizes a company would have been cleaner, it's two years of back taxes and penalties.
- A co-founder shows up. Someone joins six months in. With no entity in place, there is no clean way to assign equity, no operating agreement, no IP transfer. Whatever the two of them build is legally entangled.
None of this is hypothetical. It is the standard failure mode for indie products that grow faster than the founder expected — which, with AI in the loop, is now most of them.
"I'm too small to worry about this"
The most common objection, and wrong on two counts.
Legally, "small" is irrelevant. Liability does not scale with MRR. A free user with a copyright complaint can sue exactly like a paying enterprise customer. The threshold for being dragged into court has nothing to do with revenue.
Financially, the math is upside down. Founders delay forming a company to avoid spending $300–500. They then spend $5,000+ later, in lawyers and accountants, retroactively cleaning up two years of personal income, commingled funds, and undocumented IP. "I'll set it up when I'm bigger" is the most expensive path available.
There's a legal layer on top of that. In most jurisdictions, registration is mandatory once you cross a trivial revenue threshold (in France: €0 if the activity is habitual; in the UK: from the moment you trade; in most US states: from the first dollar of business income). Operating as an unregistered individual past that point is not a gray area. It's tax evasion.
Why an LLC specifically
For most solo founders shipping AI-built software, a US LLC is the cleanest answer:
- Liability shield from day one. The LLC is a separate legal person. Lawsuits and creditor claims hit the company, not the founder personally — provided the corporate form is respected (no commingled funds, accounts in the LLC's name, contracts signed on the LLC's behalf).
- Pass-through taxation. No corporate tax layer at the LLC level. Profits flow to the owner and are taxed where the owner lives. No double layer.
- Stripe, Mercury, Wise, AWS all open accounts in the LLC's name without friction. The retail bank ToS violation goes away.
- Clean IP ownership. The LLC owns the code, the domain, the trademarks. When the time comes to sell, raise, or bring in a partner, there is one entity holding the assets.
- Credibility. "MyApp LLC" on a contract reads differently than a personal name and a Gmail address.
A UK Ltd works for similar reasons in Europe. A Paraguayan SAS makes sense for territorial-tax optimization paired with a Paraguayan residency. The right structure depends on where you live and where the customers are. The question is never whether to have one. It's which.
What "setting it up" actually looks like
The reason most indie founders skip this is they imagine a months-long bureaucratic ordeal. It isn't, but it isn't trivially fast either, and the bottleneck is rarely where they expect.
For a US LLC formed by a non-resident, the sequence is:
- Pick a state. Wyoming for most cases. Delaware if VC fundraising is on the roadmap in the next two years.
- File Articles of Organization. Online via a registered agent. Certificate back in 1–3 days in most states.
- Get an EIN from the IRS. 4 to 8 weeks via Form SS-4 by fax for non-residents. A few days through an expedited service that calls the IRS directly.
- Open business banking. Mercury or Relay, fully online, KYC done from a phone. Plan for a possible second attempt if the first application is declined.
- Sign a one-page operating agreement. Template is fine for a single-member LLC. Useful even when there's only one owner — banks ask for it, and it's the document that proves the company is actually run as a company.
End-to-end timeline: 3 to 6 weeks for a smooth case. First-year cost is in the hundreds to low four figures, including state fees, registered agent, and EIN service. Annual recurring cost after that is materially lower, but non-zero — state report, registered agent renewal, and the IRS filings every non-resident single-member LLC owes by April 15.
That's less than what most vibe-coding founders spend on AI subscriptions in a quarter.
The full procedure (with the friction points where things actually go wrong — EIN delays, Mercury declines, BOI report at FinCEN) is in How to Open a U.S. LLC Fast.
The move, in plain terms
Open a US LLC. That's the entire intervention.
Past a certain point, doing it yourself is a false economy — the IRS forms, the BOI report, the registered agent, the Mercury KYC, the migration of an existing Stripe to a new entity all take time and have one or two non-obvious failure modes each. The cleanest path is to delegate the whole package to a provider that does this every week. That's the business we run at Leasum.
The previous revenue, however, stays on the founder's personal tax record. There is no way to retroactively shift income that already landed in a personal account. That's part of why waiting "until things get bigger" is the costly play — every month of delay is another month of revenue stuck on the wrong side of the line.
The calculus
Vibe coding has collapsed the cost of building software to near zero. That is genuinely good. It has also produced a wave of founders who treat the legal layer as optional, because the technical layer became cheap.
The legal layer is not optional. It's the part that determines whether the product is yours — protected, transferable, defensible — or a side project running on a personal name with no recourse the day something goes wrong. The cost of getting it right at the start is a rounding error. The cost of getting it wrong is the kind of thing that takes a year to unwind.
This is the gap Leasum closes. We form US LLCs end-to-end for non-resident founders — state filing, EIN, BOI report, registered agent, Mercury onboarding — and carry the annual compliance (state report, Form 5472, registered agent renewal) after that. Most founders are operational within three to six weeks, and the migration from a personal-account setup is part of the engagement: we get the LLC live, then guide the Stripe entity change, the banking switch, and the IP transfer so nothing breaks in production.
If you've shipped a product this year and it's still running on a personal account, that's the gap to close before the next $10k in revenue lands in the wrong place.
