Well, that’s exciting.
As far as I can tell, the opening is yesterday’s Claude Code release. It added a flag called --channels, described as a research preview that lets MCP servers push messages into your session. The docs put it plainly: a channel is “an MCP server that pushes events into your running Claude Code session, so Claude can react to things that happen while you’re not at the terminal.” Telegram and Discord are among the channels in the preview. Send a message from your phone, and the session running on your own machine picks it up, does the work against your real files, and answers in the same chat.
Now — can we figure out a way to manage multiple Claude Code instances through this opening? Picture several primary agents, each tasked with a specific part of your daily workflow, each reachable from a chat window. That’s effectively an open box version of OpenClaw, the self-hosted assistant whose docs describe it as “your AI assistant, on your own hardware, in every chat app you already use.”
I think it still lacks the overarching orchestration layer, though. As far as I can see, a channel gets a message into one session. Nothing yet decides which session should get it, or notices when two of them are working at cross purposes. That’s the manager’s job, and the manager is the part that’s still missing. OpenClaw’s answer was to make the hub itself the product. Claude Code, so far, has left the hub for someone else to build. Just musing.
Certainly, if you can manage multiple Claude Code instances, you can advance a specific project more effectively than OpenClaw could. OpenClaw is built to be everywhere at once — Discord, Signal, WhatsApp, Teams and the rest — with isolated sessions per agent, workspace or sender. Claude Code is built to sit inside one repository and finish the job.
So both me and my OpenClaw will be using this today. I’ll let you know what I figure out.
