Video yükleniyor...
Video Yüklenemedi
We've also added multi-root workspaces for cross-repo changes. A single agent session can now target a reusable workspace made of multiple folders.
35,681 görüntüleme • 5 ay önce •via X (Twitter)
18 Yorum

Another way to parallelize work is with new and improved worktrees in the agents window. Run isolated tasks in the background across different branches. When you're ready to test changes, move any branch into your local foreground with one click.

Introducing /multitask in the new Cursor 3 interface. Cursor can now run async subagents to parallelize your requests instead of adding them to the queue. For already queued messages, you can ask Cursor to multitask on them instead of waiting for the current run to finish.

Download Cursor 3.2 to try these new features in the agents window:

solid

How does indexing work here?

multi-root helps, but stale state between repos is the thing that still burns me. are you re-indexing both roots before each write, or trusting the initial pass?

love the multi-root workspaces thingy. Very useful for backend/frontend work or refactors from old to new codebases. This solves a real issue for me 😍

Very Exciting features. May I have some credits to experience ?

大佬 考虑给个额度

Dear Team I'm Usama From URAM Agency Working in Guest Posting/ Link Building We have many high Authority sites where I can publish guest posts. For example: 62 million traffic DA & DR 90+

cursor目前太过专注于程序员群里,但有没有想过其实cursor这类工具最大的使用人群增长点是那些本来不会写代码但是因为有了AI大模型而开始尝试做代码产品的群体呢?这类群体反而更依赖cursor的agent能力

This feature doesn't seem to support remote SSH repos.

This solves a loop I keep hitting. Agent works on the frontend, jumps to the backend functions, and by then has forgotten the response shape it set up a few minutes earlier. One workspace holding both at once fixes the hand-off.

How to get the credits?

Multi-root workspaces feel like one of those small changes that unlock much more agentic work in practice. The next step is carrying that same context across tools - code, docs, browser tasks, memory, and scheduled workflows. That’s the kind of broader agent surface we’ve been building with Thoth:

Finaly! I was missing this.

How is the .cursor stuff managed when having multi root workspaces? Which AGENTS.md takes priority for example?

Does Cursor natively allow a single prompt that will plan with one agent and build with another?
