Anyone else having a terrible experience with Claude Code's remote control?

[leddo]

It's constantly going down. I'll kick off a task, walk away, come back and the connection dropped with no way to recover the session. Yesterday it was completely unavailable for hours.

And even when it works, you can't see anything visual. If I ask Claude to build a UI component, I have no way to see what it actually looks like on my machine.

I'm flying blind, approving changes to a button I've never seen.

Am I the only one who feels like remote AI coding should be way further along than this?

[RationPhantoms]

I have not tried it but prior to them releasing that feature, I used an iOS terminal with SSH mosh capabilities + Tailscale + Tmux session mounted to see what the output of my home machine's Claude conversation was doing.

[leddo]

yeah I'm doing that too now! Thanks man!

[simonhfrost]

Yes. My chats don't get new messages without exiting and entering, I'm assuming because it's still in research preview...

[leddo]

Possibly! I'm using tarsy.dev for that now