Feedback wanted: what works, what is unclear, what breaks?

[xxwink]

Smeldr keeps decisions coherent when people and AI agents move fast. It is open source (AGPL) and self-hosted. We would like to hear what you think: what works, what is unclear, what breaks, and what you would like it to do.

What to try, in about three minutes: the quick start in the README. Run the server, make a token for yourself and one for an AI agent, ask the agent to propose a Decision, and watch it be refused when it tries to ratify.

We have run the quick start on Linux and on Windows 11 (PowerShell and Git Bash). We have not run it on macOS, so a report from there is especially welcome.

Tell us here, as short or as long as you like:

• What worked, or made you think "that is how it should be"?
• Where did you get stuck, or what did you have to guess?
• What did you expect to happen that did not?
• What would you want it to do that it does not? Feature requests are welcome.

A plain "this does not make sense to me" is useful. So is "I would not use this", with the reason. And so is "this is exactly what we need": tell us what for.

For a bug, open an issue. For everything else, answer here. We read every reply.