The cofounder breakup playbook for early stage projects

I have watched more early stage projects die from cofounder breakups than from product market fit. The market is patient. Two people pretending to still be aligned are not.

This is the playbook I wish someone had handed me the first time I was on the wrong end of one. It does not replace a lawyer or sitting with how you actually feel. But it is the shape of the thing.

The early signs it is not working

Nobody wakes up and decides to fire their cofounder. The friction builds. By the time you are searching “how to exit cofounder agreement” at 2am, the signals have been there for months.

What I see most often:

Different pace. One of you ships on weekends, the other ships when the mood strikes. Neither is wrong alone. Together, they corrode trust fast.
Different ambition level. One wants a $10M ARR business, the other wants a nice $5k MRR side income. Both are legitimate. Neither is compatible inside the same cap table.
Different commitment. Full time on one side, evenings and weekends on the other, with equity frozen at 50/50. The math feels worse every month.
One person carrying disproportionate weight. When the silent founder is privately tallying their hours against the other’s, and the gap is 3x or 5x, resentment is already living rent free in their head.
Communication breakdown. Slack messages get shorter, calls get rescheduled, the shared doc has not been opened in 6 weeks. You are roommates who barely talk.

If you recognize three of those, you are not in a rough patch. You are in the early phase of a breakup. The kindest move is to name it.

The conversation itself

The biggest mistake people make is opening with “you are not doing X.” That is a fight, and you will lose it even if you win it.

Lead with “this is not working for me.” That phrasing is honest, unarguable, and puts the responsibility where it belongs.

A version that has worked for me:

I have been thinking about this for a while. The way we are working together is not working for me anymore. I do not think it is anyone’s fault. I think we want different things, and I would rather name that now than pretend for another 6 months. Can we talk about what a clean ending looks like?

Then shut up and let them respond. Most people, when given a way out that does not involve being attacked, take it gratefully.

If they push back, do not litigate the past. The hours, the missed standups, the feature that never shipped, none of that matters now. The goal is to end a partnership well.

The legal mechanics

If you have a proper agreement in place, this is mostly paperwork.

Vesting cliffs are your friend. If your cofounder is inside their 1 year cliff and leaves, their shares typically revert to the company entirely. Not punishment, the system working as designed. Vesting exists so equity reflects work actually done.

Repurchase clauses let the company buy back unvested (and sometimes vested) shares at a predetermined price, often the original purchase price plus interest. Read your shareholders’ agreement, the clause is usually there.

The founder share buyback. If your cofounder has earned 25 percent vested, you have three real options. Buy them out at a fair valuation. Let them keep the vested portion as a passive holder, which rarely ends well. Or negotiate a smaller buyback in exchange for a clean release.

If there is a real cap table, get a lawyer. A 2 hour consultation costs less than cleaning up a sloppy split.

What if there is no agreement

In indie founder land, this is the norm. Two friends, a Notion doc, a shared Stripe account, vibes. No vesting, no operating agreement, no cap table.

You can still end this cleanly. Write the document you should have written at the start, except now it is a termination agreement instead of a founders’ agreement.

What it should cover:

Who owns the IP and code from this point forward
Who keeps the domain, the social handles, the Stripe account, the customer list
A revoke list: GitHub, Vercel, Linear, 1Password, Notion, Slack, every tool. Remove access on the call.
A non-compete clause, or more realistically, “you can build whatever you want, we just will not coordinate”
A mutual release, so neither party can come back in 18 months claiming they are owed something.

Sign it. Even a one page Google Doc with two e-signatures is meaningfully better than a handshake.

Splitting the IP and code

If one cofounder built 90 percent of the product, the IP question is mostly settled by reality. They keep building, the other person walks, a token cash payment for the early contribution can smooth the exit.

If the work was actually evenly split and you both want to keep going separately, you have a problem. Forking the codebase is technically easy and emotionally a disaster. I have seen exactly one fork end well, and only because the two markets were genuinely non-overlapping. Usually one party has to step away from the code.

What to tell users and customers

Less than you think. They do not need your internal politics.

If both of you were public faces, post a brief, calm announcement. “X is moving on, the project continues, here is who to reach for support.” That is the whole script. No recriminations. Customers care that the product still works on Monday.

If only one of you was the public face, you may not need to say anything.

Three real exit paths

Once the breakup is named and the legal piece is moving, there are three doors.

One of you buys the other out and keeps building. Cleanest if there is genuine appetite to continue solo.
You both walk away and let the project die quietly. Sometimes this is the right call. The project served its purpose, which was teaching you both something.
You list it on Failedups together, split the proceeds per your termination agreement, and let someone with fresh energy take it forward. I am biased here, but I think this is often the most respectful ending. What you built gets a second user.

The emotional reality

Most cofounder breakups follow the same arc. Two or three weeks of grief, even when you were the one who initiated it. Sleep gets weird, you second guess yourself, you re-read old Slack threads at midnight looking for the moment it went sideways.

Then, almost always, both people feel lighter. The weight you were both carrying in silence turns out to have been most of the exhaustion.

A clean breakup is far better than a slow death. The slow death takes a year, costs you the friendship, and usually kills the project anyway. The clean break gives you both a chance to do good work again.

If passing it on is the right ending, list it on Failedups when you are ready. The project deserves a second life, and you deserve to put it down.