Pluto Door
The pitch is so short it could fit on a sticky note, and honestly that's part of the appeal. An SSH client with an AI agent and a desktop for every server is one of those ideas that sounds obvious the second someone says it, then you realize nobody has shipped it well. I've spent hours bouncing between a terminal, a browser tab for metrics, and a chat window asking why something broke. Pluto Door wants to collapse all of that into one surface, and I'm here for the ambition.
What I'd really want to know is what the desktop actually shows. Logs, files, process trees, maybe a visual editor? And how smart is the agent beyond generating commands? If it can diagnose a spike in CPU usage on its own and surface the culprit, that's meaningful. If it just autocompletes ssh flags, that's a nice feature, not a new category.
The bigger question is whether any of this gets in the way. Terminal purists are a stubborn bunch, and the fastest path to rejection is adding friction to a workflow people have muscle-memorized. But the era of AI agents reading systems and taking actions is coming, and the terminal is the natural home for it. Pluto Door is betting on that faster than the big cloud providers, and I'd love to see it deliver.