Yeah, totally understand keeping it simple... and visualizations should probably be externally injected via plugin anyway to keep payload smaller where not needed.
Webamp uses butterchurn, for reference.
https://github.com/jberg/butterchurn
Pretty nice overall... semi-mixed feelings about the css customization vs a theme property, but it is closer to the browser's intended use for this.
Would be cool to see this extended with some visualization support like the webamp.org player.
It's interesting... First look and the example todo app are pretty easily understood... I'm still not sure how I feel about decorators defining UI integrations, though it appears at least at first glance much better than say Angular's approach. It's been a while since I looked at Vue, so don't recall it well enough to draw comparisons there.
Generally speaking, however, I've been really happy with the react-style component approach, especially with hooks and functional components. I've taken to using more of an MVC style approach with a controller to handle events and generate actions, then passing state change actions to a reducer (redux). It's not quite the prescriptive approach, but has made state very testable.
All of that said, it's nice to see new approaches... I'm curious how the yaml front matter in the .mahal files plays into things, since I didn't try running the application, but if it integrates with some sort of in-browser tooling extension, that could be very cool.
See fclone in npm for a version that overcomes some of the structuredClone limitations.
If you want something for deno, I created a version for that.
https://github.com/tracker1/deno-lib
It wasn't particularly click-baity, the all caps (and when people add emojis) is unwelcome by most.
The main point was that lists that don't have an associated method of *HOW* that list was generated, or why the said items are on that list etc, or mention of what was/wasn't included is significant.
When mentioning speed/performance/fast, it often comes down to metrics. I've personally enjoyed your posts... I didn't downvote myself, I'm only speaking to why someone might have downvoted... And to me it largely comes down to "Top N for X" type lists that don't have any methodology are often deleted outright... framework/tool comparisons in particular are often low quality/listicles... I left this in as there's value to it. I can't control how others vote.