Connect a repo to the IDE
Pick which workspace root the IDE opens, and switch between fleet repos.
AI
Preview — early access.
Overview
The IDE reads and writes against exactly one workspace root at a time. "Connecting a repo" means picking which one that is — either at first load, or any time after from the header.
How it works
The IDE asks the server for the available workspace candidates (each with a label and an absolute path, one marked as the fleet default). If you haven't explicitly picked one yet and more than one candidate exists, it opens a picker automatically on first load. Switching workspaces resets the open file tree and tabs — pinned chat sessions aren't tied to a workspace, but unsaved file edits are discarded after a confirmation.
Steps
- •
Open the IDE
If this is your first time and more than one workspace is available, a picker opens automatically.
- •
Pick a workspace
Choose the repo you want to work in — the fleet default is marked, and your current choice is marked "current."
- •
Confirm you're in the right one
Check the 📁 chip in the IDE header — it always shows the active workspace's label.
- •
Switch later, anytime
Click the 📁 chip to reopen the picker and switch to a different workspace root.
- •
Save first
If you have unsaved changes, switching prompts a confirmation — save first if you want to keep them.
Capabilities
- Multiple workspace candidates — pick between fleet repos rather than being locked to one.
- A remembered explicit choice — once you pick, the IDE won't re-prompt you every load.
- Unsaved-change guard — switching with dirty tabs asks for confirmation first.
Limits & honest scope
- The IDE opens one workspace root at a time — there's no split multi-repo view.
- Available candidates are whatever the server exposes for your account; adding a brand-new repo to the candidate list is a workspace-configuration change, not a self-serve "add repo" button.