Claude Code sessions on macOS and Linux can now send messages to each other and share context across terminals. The update, documented by Anthropic on August 8, lets one Claude Code instance fire off a text summary to another session, eliminating the manual copy-paste loop developers hit when running parallel tasks.
The feature works bidirectionally. You can query another session and receive an answer directly. Claude can also push messages unprompted, say, when a code change in one terminal affects what another session is working on. Anthropic's documentation lists typical use cases: sharing insights, coordinating parallel workflows, and checking status updates on long-running tasks.
How session communication works
Claude Code Sessions Can Now Message Each Other
On a single machine, communication stays local. Between different computers, messages route through Anthropic's servers, though only responses are possible in that cross-machine scenario. The distinction matters for teams thinking about latency and data residency.
Admins can lock down the feature via settings. Anthropic's documentation explicitly notes it is not available on Amazon Bedrock, Google Cloud Agent Platform, or Microsoft Foundry. If your Claude Code deployment runs through one of those managed services, you're out of luck for now.
What this changes for multi-agent workflows
The practical payoff is clearest for developers running multiple Claude Code instances on the same project. One session can handle test generation while another refactors a module. Previously, keeping them in sync meant copying context by hand. Now Claude handles the handoff.
This is a step toward the multi-agent orchestration pattern that's been gaining traction in AI engineering circles. Instead of a monolithic prompt that tries to do everything, you split work across specialized agents and let them coordinate. The hard part has always been the plumbing. Anthropic is building that plumbing into Claude Code itself.
Compares Claude Code's performance against competing AI coding tools.
Platform restrictions limit reach
The exclusion of Bedrock, Google Cloud Agent Platform, and Microsoft Foundry is worth noting. Enterprises running Claude through those managed channels cannot use session communication today. Anthropic did not say whether or when support would extend to those platforms.
For teams already on direct Anthropic access with macOS or Linux, the feature is live now. Windows support was not mentioned in the documentation.
Logicity's Take
This is infrastructure, not flash. Session communication turns Claude Code from a single-threaded assistant into something closer to a team you can dispatch in parallel. For AI builders running complex codebases, it cuts the coordination tax on multi-agent setups. The managed-platform exclusions are the main caveat: if you're on Bedrock or GCP, you're still waiting. Competitors like Cursor and Codeium don't yet offer native agent-to-agent messaging, so Anthropic has a temporary lead here.
What comes next
The open question is how far Anthropic pushes this. Cross-machine messaging already routes through their servers. A natural extension would be persistent session state, letting Claude Code resume context after a terminal closes. That would make multi-day coding tasks more practical.
For now, the feature fills a specific gap: keeping multiple Claude Code instances aware of each other without manual intervention. Whether that gap mattered to your workflow is the test.
Need Help Implementing This?
If you're building multi-agent workflows with Claude Code and want help designing your orchestration layer, reach out to the Logicity team for a consultation.
Source: The Decoder / Matthias Bastian
Manaal Khan
Tech & Innovation Writer
Produced with AI assistance and reviewed by the Logicity editorial team. Learn more in our Editorial Policy.
