Claude Code Agent View: single-screen management concurrent sessions

Anthropic on May 11 introduced a new feature for Claude Code called “Agent View,” which consolidates multiple simultaneously running Claude Code work sessions into a single screen for management, eliminating the need to switch back and forth between multiple terminal tabs. According to Anthropic’s official blog, the feature is rolling out in a Research Preview format and is available for Pro, Max, Team, Enterprise, and Claude API solutions. A single post on the official X account has received more than 18k likes, making it the highest-signal product update from that account in recent days.

Agent View interface: a list that manages all sessions in one place

Agent View provides a centralized dashboard that lists all active Claude Code sessions. Each row shows four pieces of information: the session identifier, whether it’s waiting for user input, a preview of the most recent model response content, and a timestamp for the last interaction.

Interaction methods:

Entry: In Claude Code, press the left key or run the claude agents command

Inline replies: You can directly “peek” at the previous round of conversation content from the list row and respond on the spot; there’s no need to switch back to that session

Background execution: In the existing session, switch to the background using the /bg command; or use claude --bg [task] to start a new task directly in background mode

Status tracking: The list clearly labels three states: “running,” “waiting for input,” and “completed”

In the announcement, Anthropic lists three typical use cases: first, dispatch multiple idea tests in parallel and finally return to a “PR checklist ready for code review”; second, manage long-running agents (e.g., PR monitoring, dashboard auto-updates), and use Agent View to see the next scheduled execution time; third, while conducting core development in the main session, handle tasks in a background agent so they don’t affect the main workflow.

Connections to other Claude Code updates this week

After the May 6 Code w/ Claude developer conference, Anthropic continued releasing Claude Code series features. Agent View takes the “Multi-Agent Layer” from the six-layer architecture released to the public last week and turns it into a built-in visual management interface within the CLI.

Within that six-layer architecture, the Multi-Agent Layer already supports multi-agent collaboration mechanisms such as subagent spawner, teammate mailboxes, and worktree isolation. Agent View turns the collaboration that originally happened in code into a list that “humans can see and intervene in.” Combined with the Claude Code mobile push announcement from the same day, the “monitoring and intervention” experience for multi-agent long-running tasks is gradually completed.

What to watch next

The functionality in the Research Preview stage may continue to be adjusted, and Anthropic has not disclosed an official GA timeline. From a product perspective, Agent View completes Claude Code’s image shift from a “single-session CLI” to a “multi-agent collaborative workstation,” representing a differentiating move by Anthropic in the competitive landscape versus other AI coding agent products (such as OpenAI Codex, Cursor, Cline).

Events that can be tracked next include: whether Agent View expands to other Claude Code surface interfaces like VS Code, JetBrains, and the Web; whether the community believes that “multiple agents running simultaneously” brings new workflows (e.g., a collaborative PR review agent); and whether Agent View affects Anthropic’s token consumption pattern (parallel sessions typically mean total token usage expands). For readers who want to dive deeper into Claude Code’s full capabilities, they can refer to abmedia’s complete Claude Code guide.

This article, Claude Code Agent View: Single-Screen Management of Parallel Sessions, first appeared on the news site ABMedia via Chain News.