Shipping an MVP on Base in eight weeks (and what we cut)
What an honest 8-week mainnet plan looks like — the things you build, the things you defer, and the audits you actually need.
When a founder says 'we want to launch on mainnet in two months', most studios either nod and lie, or quote a roadmap that's optimised for their margin.
We've shipped enough MVPs on Base, Optimism and Arbitrum to know what fits in eight weeks and what does not — and the difference comes down to one thing: ruthless scope.
An eight-week mainnet plan is roughly four weeks of contracts, two weeks of frontend and indexing, one week of audit remediation, and one week of buffer. Anything that does not fit in that envelope is a v2.
What goes in: the core protocol, a thin frontend, a Subgraph or a Postgres mirror, paymaster-sponsored gas where it changes the user journey, and a Foundry test suite with invariants.
What gets cut: governance v1, a points programme, a referral system, a cross-chain bridge, a token. Every one of these is a project of its own. We schedule them, we don't ship them.
The audit is non-negotiable. Even an internal review by a second senior engineer catches the class of bug that costs you the launch.