Cursor Agents Window: Multi-Agent UI, Cloud Runs & PR Handoff

You switch between the two from the Command Palette, and you can switch back at any time.

Cursor's classic editor is a fork of VS Code. For Cursor 3An older release label for Cursor's multi-agent interface; current product guidance calls the active surface the Agents Window. Press Enter for the full definition., the company went a step further and built this interface from scratch, centered around agents. One Cursor field engineer described the shift as deliberate: the company has moved away from the file-tree-and-debugger view and put the agent at the center. That matters most for designers, PMs and other non-engineers who never wanted the IDE surface in the first place.

The layout has three regions. You start agents from the prompt window in the center. The agent pane at the top left holds several workspaces or repositories in one instance and runs agents across all of them, and each workspace keeps its own index. The panel on the right is for inspecting the work, the way you would in an IDE.

Cloud agents appear here too, and you can start them from the same window. Cursor's docs list remote SSH among the environments the Agents WindowCursor's surface listing your agent runs; open a run to read its diff and search transcripts from the command palette. Press Enter for the full definition. works in. In a Cursor webinar on mobile development, the presenter said SSH in the Agents Window had been updated to be "a little bit more functional for remote development than the IDE". Try it if your target is a remote or embedded platform you reach over SSH.

Agents can review their own work in the review pane, and you can branch and commit from the same place. The current branch shows at the top right and on the review tab, which sits at the top left once there are changes. Next to the commit-and-push button is a small arrow.

Yes, by running it as a sidecar. In Cursor's webinar on mobile development, the presenter ran the Agents WindowCursor's surface listing your agent runs; open a run to read its diff and search transcripts from the command palette. Press Enter for the full definition. next to Xcode or Android Studio and moved back and forth between them. The work flows in both directions, which is how the Agents Window covers IDEs Cursor can't drive natively.

The other route is ACPAgent Client Protocol. An open standard that lets any IDE host any coding agent; Cursor uses it to run inside JetBrains/IntelliJ and Android Studio. Press Enter for the full definition. (the Agent Client ProtocolAn open standard that lets any IDE host any coding agent; Cursor uses it to run inside JetBrains/IntelliJ and Android Studio via the Cursor CLI bridge. Press Enter for the full definition.), which runs Cursor's agent inside the other IDE's own chat instead of beside it. Cursor documents ACP for JetBrains IDEs, and the presenter described it as one option for Android Studio. ACP also reaches well beyond JetBrains. The presenter pointed to open-source ACP integration work for Neovim and Emacs, and to Obsidian, where Cursor can act as an agent inside your note editor while you write.

Because it is built for running many agents at once, the Agents WindowCursor's surface listing your agent runs; open a run to read its diff and search transcripts from the command palette. Press Enter for the full definition. has features the editor lacks. You need them once agents write most of the code and your job shifts to managing the runs.

That change reaches past engineering. A Cursor support engineer described starting one agent per incoming ticket, such as "how do I log in with SCIMSystem for Cross-domain Identity Management. A standard for automatically creating and removing user accounts when people join or leave. Press Enter for the full definition." or "what does the analytics dashboard show". The engineer reviews the work as it streams in and opens each agent individually, with little context switching. Time is the scarcest resource in support, and this setup lets one person keep up with a steady flow of tickets without waiting on a single agent.

The changelog that introduced the pull-request handoff called it /babysit. Cursor's docs now name it /autopilot. The three cloud rows need a plan that includes Cloud AgentsAgents that run in a Cursor-managed virtual machine, check out the repo, do the work and open a pull request, then shut down, with no load on your laptop. Press Enter for the full definition., which start at Premium on The X SubscriptionThe subscription Cursor is sold through: one plan for Grok, Cursor and X, with Free, Plus, Premium, Super and Ultra tiers, per-seat Teams, a weekly included pool for Cursor models shared with Grok, and third-party models billed at API list price from prepaid credit. Press Enter for the full definition..

AI coding has moved through three eras. Tab predicted your next edit while you typed. Synchronous agents then edited across files while you watched. Now asynchronous cloud runs work toward a goal on their own branch. Day to day, the choice that matters is between the last two: keep a task in the foreground when it needs your judgment, and send it to the background when its output can prove itself.

Foreground is for decisions: the task is ambiguous, it needs steering, you want quick front-end feedback, you are watching it live, or it touches something security-sensitive. You see and control every step. Background is for execution against a clear goal with acceptance criteria, where the output is artifact-verifiable: an image, a recording, a log or a diff you can read after the fact.

Stay in the editor when you want VS Code extensions, flexible split panes and manual navigation across many open files. Use the Agents WindowCursor's surface listing your agent runs; open a run to read its diff and search transcripts from the command palette. Press Enter for the full definition. when agents write most of the code and your work is starting and reviewing parallel runs. Cursor's docs say the Agents Window pulls you up to a higher level of abstraction, and that Cursor will keep supporting and improving both.

The choice isn't permanent. You can switch at any time or keep both open. For a quick look at one file you don't need to switch at all, because ⌘+P and ⌘+⇧+F work inside the Agents WindowCursor's surface listing your agent runs; open a run to read its diff and search transcripts from the command palette. Press Enter for the full definition..

Agents WindowCursor's surface listing your agent runs; open a run to read its diff and search transcripts from the command palette. Press Enter for the full definition. became generally available with Cursor 3An older release label for Cursor's multi-agent interface; current product guidance calls the active surface the Agents Window. Press Enter for the full definition., released on April 2, 2026. For the two weeks after launch, Enterprise admins could give access to their whole team or to specific users from Team settings. After that rollout period, Cursor's docs say all users have access by default.

Cursor's Agents WindowCursor's surface listing your agent runs; open a run to read its diff and search transcripts from the command palette. Press Enter for the full definition. docs don't describe a control for limiting access after the rollout ended. If your organization needs one, confirm what your admin dashboard offers before you plan around it.