Claude Code Agent View: Manage All Your Parallel Sessions in One Unified Interface
Claude CodeAnthropic launched agent view in Claude Code as a Research Preview on May 11, 2026, giving developers a unified dashboard to oversee all their parallel Claude Code sessions from a single screen. Accessed via claude agents or the left arrow key from any session, the interface shows each session's status, last response, and whether it needs input — eliminating the need to juggle multiple terminal windows. Users can peek at a session, reply inline to unblock a waiting agent, or attach directly for full transcript navigation.
Sources & Mentions
5 external resources covering this update
Manage multiple agents with agent view
Claude Code Docs
Thariq (Claude Code team) on agent view design
X
Anthropic adds Agent View to Claude Code CLI interface
Testing Catalog
Claude Code introduces specialized sub-agents
Hacker News
Claude Code's 'Tasks' update lets agents work longer and coordinate across sessions
VentureBeat
Claude Code Agent View: One Screen to Rule All Your Sessions
On May 11, 2026, Anthropic introduced agent view in Claude Code — a Research Preview feature that fundamentally changes how developers manage parallel workloads. Available starting with v2.1.139, agent view consolidates every running, blocked, and completed Claude Code session into a single unified list, making multi-agent workflows feel as natural as tabbed browsing.
What Agent View Looks Like
Pressing the left arrow from any active session, or running claude agents from the terminal, opens the agent view. Each row in the list displays:
- The session's current status (running, waiting for input, or completed)
- The last response content from that session
- The timestamp of the most recent interaction
Users can navigate the list with arrow keys and select any session to preview its final exchange without fully attaching.
Peek, Reply, and Attach
One of agent view's most practical capabilities is inline reply. If an agent is blocked — waiting for a decision, a file path, or a confirmation — the user can respond directly from the agent view list and the session resumes automatically, without switching contexts. For deeper exploration, pressing Enter attaches to the session for full transcript navigation.
Sending Sessions to the Background
Existing interactive sessions can be moved to agent view at any time using the /bg command. New sessions can also launch directly in background mode with claude --bg [task], making it straightforward to dispatch multiple work streams at once and return to them when results are ready.
Typical Workflows
Anthropic highlighted several patterns that early users have adopted:
- Parallel ideation: dispatching multiple concurrent tasks — each optionally paired with a Claude Code skill — and returning to find pull requests ready for review
- Long-running monitors: keeping persistent agents like PR monitors or dashboard updaters visible in agent view, with next-run time displayed
- Lightweight navigation: scanning which sessions produced pull requests or flagged issues without leaving the overview
Availability
Agent view is available as a Research Preview on Pro, Max, Team, Enterprise, and Claude API plans. Standard rate limits apply across all sessions. Each session uses subscription quota independently.