LLMs are great at generating aesthetic HTML/CSS. But there's nowhere to actually use that output. There's no code native tool to edit, present, and share decks built as HTML. So we built one. Variant is a presentation editor where slides are real HTML documents. Agents author them directly (Claude Code or Codex via MCP works great), then you immediately open that deck in a presentation viewer/editor/collaborator (think Google Slides) and finalize the deck by hand, then present. Models are best with code, they aren't great with working with elements like in powerpoint, and even image generation has a ton of pain points. HTML lets it reason in flex/grid and components, which it's great at. Plus you get embeds, SVG, canvas, and interactive widgets in slides for free. Great writeup on model generated HTML here: x.com/trq212/status/2052809885763747935 Happy to answer questions or hear feedback.

(Edit) nvm, the usual Xcancel (xcancel.com/trq212/status/2052809885763…) just links to an article (x.com/i/article/2052796100608974848)

I’ve done a lot of exploration here. Visual explainers certainly come in hand. You need to have both the extra time and cost of tokens accounted for when you ask for these. As well as maintainability - you lose things like vcs/diffs over plans that are much more readable as markdown. The more interactivity on the surface, the better. Esp if they create feedback loops with the agent. Simple annotations to start. Markdown still goes a long way. I’ve effectively been enriching plan rendering in Plannotator. Now supporting GFM, custom svgs, etc. github.com/backnotprop/plannotator For fun, I also enabled annotations on top of these type of svgs, see here x.com/plannotator/status/20528517319989…

If we want to close Human Machine loop as much as possible(pre-neuralink). Assuming that today the most efficient way for human to transfer information to machines is via voice. Assuming for machines to convey rich information to humans that's by printing html. Then a combination of screen + eye tracking + voice is all you need. The mouse doesn't make sense anymore. Links: x.com/trq212/status/2052809885763747935