Made by someone with
eleven unfinished projects.

Wanju started the way most of its users' projects did: as a side project. The difference is what it was for — a way to look at a folder of half-built apps and research experiments without flinching.

The insight behind it is simple. AI coding agents made starting effortless, which made the graveyard grow faster. What nobody built was the other half: a system for finishing — knowing the true state of each project, the single next step, and a way to delegate that step without giving up control of your own code.

So Wanju is opinionated in three ways. It's portfolio-first: the unit of work is your whole folder of projects, not one repo. It's deny-by-default where Wanju controls the action: the built-in agent queues proposals for approval, while external CLIs keep their own permissions and are reviewed in Source Control. And it's local-first: your code, keys, and audit trail live on your machine, because trust you can't inspect isn't trust.

The name is Korean: 완주 (wanju) — "completing the full course." It's the word for crossing a marathon's finish line, not just entering the race. That's the entire product philosophy in two syllables.

Questions, corrections to the comparison page, or just want to show off something you finally finished? Write to hello@wanju.dev — that's the fastest way to reach us.