I hadn't seen herdr until just now but I built a similar tool with it's own keyboard bindings https://github.com/hadlock/cscope . Claude Code changed some underlying changes to copy-paste so it needs updating. But I think this kind of "I wish my workspace did this/had these features" is a pretty common trope. It is about $10 worth of your employer's monthly token allocation (or personal account if you pay for one) and then improves output so I call it a win-win. Mine is a bit more chaotic, 80s era neon color scheme but accomplishes the same feature. All those "personal tooling" blog posts get voted up for a reason.

TL;DR why would I accept someone elses' vibe coded tool, when I can build my own to my exact specification?

Yes, absolutely inspired by herdr. The main reason I built hrdx was that I wanted to explore the idea in Go and shape it around my own workflow. I focused on persistent PTY sessions, workspace and Git branch organization, agent activity indicators, custom agent harnesses, and a socket API for external automation. There is definitely overlap, and I'm not claiming the underlying idea is unique. It is mostly my own take on the same problem.

True. The features arent groundbreaking.

But fair question. If tmux already fits your workflow, there may be no reason to switch. hrdx is more opinionated around running multiple coding agents across projects: it provides a workspace and pane sidebar, agent activity and completion status, built-in launch/resume flows, persistent sessions and layouts, and an external control API. The tradeoff is that it is much younger and less configurable than tmux.

Try it, play around with it, love it or hate it :-)

zmx, another multiplexer, has a good argument for this: https://bower.sh/you-might-not-need-tmux

Claude Code stopped working in tmux for me at some point. Which is of course more of a CC problem than a tmux one, but maybe OP was trying to work around such issues.