Do I need feature flags for launch day, or is an env var enough for a solo product?

The hosted services start earning their keep when you need percentage rollouts with sticky assignment, targeting rules more complicated than a list of ids, an audit trail of who flipped what, and non engineers flipping flags safely. None of those describe you in three weeks. Most of them do have free tiers that would cover a solo product, but the paid tiers are priced per seat per month in a way that assumes a team, and the real cost is that you now have an external dependency in your request path on launch day. I would not add one this month.

The mistake I made was flagging everything. Two years later I had around 60 flags, most of them permanently on, several with dead code paths behind them that no longer compiled if you flipped them, and a test suite that only exercised one combination. Cleaning it up took a full week. Put an expiry date in the table row when you create a flag, even if you ignore it, because it makes the graveyard visible.

A flags table plus a 30 second cache is about 40 lines and it covered me for two years. Row per flag, a boolean, an optional list of account ids for early access, and a helper that reads it. The two things I would insist on from day one are that the check is a single function everywhere so you can change the backend later, and that flags default to off if the lookup fails, because the day your config query times out you do not want every feature turning itself on at once.

Config table, cached, defaults off, one helper function. Revisit when a second person can deploy.

Slightly different angle: the flag is not the safety mechanism you think it is, your rollback is. Time yourself doing a full revert to the previous deploy right now, today, before launch. If that number is under two minutes, you can turn almost anything off by deploying, and the flags are a convenience. If it is twenty minutes because your migrations are not reversible, no flag system saves you, because the thing that breaks on launch day will be a schema change and not a feature toggle.