Hello,
I’m experiencing a somewhat strange issue. In some projects, the chat interface includes a button that allows running multiple agents or AI models, while in other projects this option is completely missing.
Why does this difference occur between projects?
Is there a specific setting, feature flag, or project configuration that needs to be enabled for this option to appear?
I’d appreciate any clarification or guidance.
Thank you.
Hey, thanks for the report.
This is expected behavior. The “Use Multiple Models” feature isn’t supported in multi-root workspaces (.code-workspace files).
Can you confirm you’re opening the project where this option is missing via a .code-workspace file?
Workaround: Open Cursor directly in the repo root folder (where the .git folder is), instead of using the workspace file. This should bring the feature back.
We’re tracking the request to support this in multi-root workspaces here: Enable Multi-Agents Support for Multi-root Workspaces
Let me know if this helps!
“I hadn’t initialized Git for the project, but after I created a Git repository, the option appeared. Thank you.”
