Agent view and background agents
A screen for running and watching several Claude Code sessions at once, so long tasks proceed while you work on something else.
Last verified August 21, 2026 · 3 sources
Agent view is the dashboard: several full Claude Code sessions running in parallel, visible from one screen.
Where it sits among the parallelism features
There are three overlapping concepts and they are genuinely different:
| Feature | Unit | Purpose |
|---|---|---|
| Subagents | Child agent inside one session | Contain context cost of a subtask |
| Agent teams | Multiple sessions that message each other | Coordinate strands of one effort |
| Agent view | The interface | Watch and control several sessions |
Agent view is the window, not the orchestration. You can use it with independent sessions that have nothing to do with each other.
Worktrees usually come with it
Two sessions editing the same working directory will fight. Git worktrees give each one its own checkout, which is why Anthropic documents them alongside parallel sessions rather than as a separate git tip.
What it does not do
- It does not merge results. Combining work is still yours, or an agent team’s.
- It does not multiply your usage allowance. Parallel sessions consume in parallel.
Plan availability
Anything left blank is not documented publicly, not a claim that it is unavailable. Compare every feature at once in the availability matrix.
- FreeNot available
- ProIncluded
- MaxIncluded
- TeamIncluded
- EnterpriseIncluded
- APINot available
Sources
- 01Manage multiple agents with agent viewOfficialcode.claude.com
- 02Run agents in parallel — Claude Code docsOfficialcode.claude.com
- 03Run parallel sessions with worktreesOfficialcode.claude.com
Page changelog
- VerifiedConfirmed as the documented interface for running several full sessions in parallel.