Incorporation, high level

~8 min read

Why it matters

A company is a separate legal person that can own the product, sign contracts, open a bank account, and issue equity. Until that exists, you own the risk and the IP in a messy way, and an investor usually will not wire money. You do not need to become a lawyer. You do need to know the one US pattern venture-backed software startups actually use, when it starts to matter, and what to hand a lawyer instead of inventing.

This page is not legal advice, not a filing guide, and not a comparison of entities outside the United States. If you are not building toward a US venture-style company, stop here and talk to counsel in the place you actually operate. Country-specific incorporation is an explicit non-goal of this playbook.

Core concepts

The common US venture pattern is a Delaware C-corporation. Investors, option plans, and startup law firms are set up for it. That is a coordination fact, not a moral claim that Delaware is “best” for every student project, consultancy, or non-US team. An LLC can be the right box for a small cash business you do not plan to raise venture for. If you might raise from angels or funds, don’t improvise a structure — ask a startup lawyer early, because converting later is a project.
Separate means separate. The company should own the IP, the domain, the bank account, and the contracts. Founders assign IP in. Mixing personal Venmo, personal laptops full of unassigned code, and a “we’ll paper it later” handshake is how a small fight becomes a company that can’t be financed.
Timing is “before other people’s money and before serious IP,” not “the morning of demo day.” You can talk to users as a person. Once someone is paying real money, a cofounder is writing code you intend to own together, or an investor is interested, the entity and the assignments need to exist. Earlier is paperwork. Later is a mess.
Use a standard path. Clerky-style startup formation services, or a lawyer who does this every week, exist so you get boring, investor-familiar documents. Novelty in your charter is not a feature. You are not trying to be clever about corporate form.
Paper you should expect to exist, without memorizing it: certificate of incorporation, bylaws, founder stock purchase agreements, IP assignment, and a cap table that matches those papers. A lawyer or a standard service produces these. You read them. You do not copy a random gist.
Equity paperwork has deadlines that are not optional. Vesting and any tax election a lawyer tells you about (people will mention an 83(b) in the US) are time-sensitive. The action is: ask counsel the week the stock is granted, not after you Google it at month three. This playbook will not walk you through a filing.
Bank account, bookkeeping, and “don’t commingle” are the operating minimum. Open a company account once the entity exists. Pay company costs from it. Keep every receipt. That is the whole accounting system you need until runway gets more complicated — not an ERP.
If you are joining rather than founding, you are not incorporating. You are checking that they did: the offer comes from a company, IP you write will be assigned to it, and someone can tell you the state of the equity paperwork. Vague answers are a reason to slow down.

What to do next

Decide which sentence is true: “we may raise venture or issue options in the US” or “this is a small project / non-US and we need local advice.” Do not straddle.
If it’s the US venture path, book a startup lawyer or a standard formation service this month. Bring: founder names, roughly equal or not (see Equity), who will be an officer, and that you want a boring Delaware C-corp. Ask them what you should not do before papers are signed.
List every asset that must move into the company: repos, domains, designs, name, any code written before the company. Nothing stays “on my personal GitHub, trust me.”
After the entity exists, open the bank account and point runway tracking at it.
Write down questions for counsel instead of resolving them in a group chat: founder stock, vesting, IP, and whether anyone is being paid yet.
Do not file random forms from a blog post, and do not take entity advice from an investor’s tweet. The cost of a lawyer hour is smaller than an unfinanceable cap table.

Mental model

  you, as people                 the company (a separate box)
  ---------------                ----------------------------
  ideas, labor, reputations      owns IP, contracts, cash
                                 issues stock to founders
                                 can hire and raise

  US venture default for this box: Delaware C-corp
  (a convention investors expect — not a global answer)

  your job:  choose the path, use standard docs,
             move IP in, keep cash separate,
             ask a lawyer about deadlines
  not your job:  invent a structure, or become the lawyer

Boring corporate form is a feature. The novel thing should be the product.

Common failure modes

A semester of code with no assignment. The person who leaves still owns what they wrote. Fixing that under time pressure, or after a fight, is miserable.
“We’ll stay an LLC and convert if we raise.” Sometimes right, often a surprise tax and legal project. Decide with a lawyer before the cap table has friends on it.
Commingled money. Personal expenses out of the company, or company revenue into a personal account, blows up both the books and the “separate entity” story.
Optimizing tax trivia instead of talking to users. Entity choice does not find a customer. Spend an afternoon on formation, then go back to the wedge.

Watch

Legal and Accounting Basics for Startups with Kirsty Nathoo and Carolynn Levy (HtSaS 2014: 18) — Y Combinator. Formation, why teams use a Delaware corporation, founder paperwork, vesting, and the basics of not creating a mess. Watch it as a map of topics to ask a lawyer, not as instructions to file alone.

A second video is omitted on purpose: most “how to incorporate” material online is either a product ad or jurisdiction-specific tax advice, and this playbook will not guess URLs.