Short answer: OpenKnowledge helps agents write your docs. Stet makes sure you still read them.
On paper we look alike. Both local-first. Both plain markdown, git underneath. Both desktop apps built for people who write with agents. If you found both on the same day, the confusion is fair.
OpenKnowledge is a knowledge base built so agents can maintain it. It gives them native tools, search over your whole wiki, and concurrent editing, so your docs stay current with as little human effort as possible. For an engineering team's runbooks and internal wikis, that's a reasonable goal: the docs serve the system, and fresher is better.
Stet starts from the opposite end. Some documents don't serve a system. They carry your name. The spec your team builds from, the brief a client reads, the post you publish. For those, "the agent kept it current" isn't reassurance. It's the problem.
So where OpenKnowledge removes the human from the loop, Stet builds the product around the loop. Every change any agent makes arrives as a proposal. You review it, see exactly what moved, and accept or reject it. Nothing lands without your sign-off, and the history of what you approved stays with the file.
If you want a wiki your agents tend while you build, OpenKnowledge is a fine choice, and it's open source. If the words are the work, and someone will eventually ask "did you write this?", you want the editor that let you answer yes on purpose. Same markdown either way. You can move between them without asking anyone's permission, which is rather the point of plain files.
Part of Where Stet fits, the map of the markdown neighborhood.