Claude Code adds agent view to manage parallel sessions

Claude has just rolled out agent view for Claude Code, a CLI dashboard for tracking all your sessions in one place. It shows what’s waiting on you, what’s still running in the background, and what’s finished. Available now as a Research Preview.

claude cover

TL;DR

  • Agent view (Claude Code CLI): Central list to manage all Claude Code sessions; reduces multi-tab/tmux overhead
  • Session status tracking: Shows waiting-on-input, working, finished; sessions can run in background
  • Launch/open commands: Left-arrow from any session or run claude agents
  • Per-session details: Row shows input-needed flag, last response preview, and last interaction time
  • Inline control: Preview latest turn, reply inline to resume, Enter attaches for full transcript view
  • Background workflows: /bg backgrounds work; claude --bg [task] starts new background session; Research Preview on paid/API plans

Claude has introduced agent view in Claude Code, a CLI-based interface the company describes as a way to manage “all your Claude Code sessions” in one place.

The feature appears aimed at users running multiple agents in parallel, replacing a setup that could involve several terminal tabs, a tmux grid, and manual tracking of which task still needs attention. Claude states that agent view lets sessions run in the background while surfacing which ones are “waiting on you,” which are still working, and which have finished.

How it works

Agent view can be opened by pressing the left arrow from any session or by running claude agents in the terminal. Each row in the list shows the session, whether it needs input, the contents of the last response, and the time of the last interaction.

Users can also select a session to preview the latest turn. If a session is waiting on a decision, Claude says replies can be entered inline and the session resumes. Pressing Enter attaches directly to a session for a full transcript view.

Claude also supports moving work into the background with /bg, while claude --bg [task] can launch a fresh session without starting in the foreground.

How developers are using it

Claude lists a few use cases early users appear to be adopting:

  • Running many concurrent sessions: multiple ideas can be dispatched at once, sometimes paired with a skill, with the results later reviewed as a list of pull requests.
  • Handling long-running agents: looping jobs such as PR babysitters and dashboard updaters show their next run time in the list.
  • Switching between related sessions: users can jump into a quick codebase question, then return to the original session after the answer arrives.
  • Tracking what shipped: status indicators and the peek title can make it easier to spot which sessions produced a PR.

Availability

Agent view is available as a Research Preview on Pro, Max, Team, Enterprise, and Claude API plans. Claude says access can be enabled by running claude agents, with usual rate limits applying. Documentation is available here.

Source: Claude

Continue the conversation on Slack

Did this article spark your interest? Join our community of experts and enthusiasts to dive deeper, ask questions, and share your ideas.

Join our community