You are halfway through a feature.
Your files are modified, the code is not ready to commit, and suddenly an urgent production issue comes in.
The usual workflow is:
git stash git switch main git pull git switch -c hotfix/production-issue
Then after the fix:
git switch feature/my-feature git stash pop
It works, but repeated branch switching and stashing quickly becomes painful.
A cleaner option is Git worktrees.
What is a Git worktree?
A worktree lets you check out multiple branches from the same Git repository into separate folders.
Instead of:
one folder
↓
switch branches repeatedly
you can have:
springboot-example-svc/
→ feature/customer-search
springboot-example-svc-hotfix/
→ hotfix/invalid-customer-id
Each branch gets its own working directory.
Your uncommitted feature work stays exactly where it is.
No stash required.
Example
Assume you are working on:
cd ~/projects/springboot-example-svc git switch -c feature/customer-search
You have unfinished changes in:
CustomerController.java CustomerService.java
Then an urgent production bug arrives.
Instead of switching branches, create another worktree:
git fetch origin
git worktree add \ -b hotfix/invalid-customer-id \ ../springboot-example-svc-hotfix \ origin/main
Now you have:
springboot-example-svc/
feature/customer-search
unfinished work stays here
springboot-example-svc-hotfix/
hotfix/invalid-customer-id
clean workspace for the fix
Open the second folder in another terminal or IDE window:
cd ~/projects/springboot-example-svc-hotfix
Fix the bug, test it, commit it, and push it:
./mvnw test
git add . git commit -m "Fix invalid customer ID handling"
git push -u origin hotfix/invalid-customer-id
Raise the PR and merge it into main.
Meanwhile, your feature workspace was never touched.
Go back to it:
cd ~/projects/springboot-example-svc
You are still on:
feature/customer-search
with all your unfinished work intact.
Once the hotfix is merged, bring the latest main into your feature:
git fetch origin git rebase origin/main
or:
git merge origin/main
When you are done with the temporary workspace:
git worktree remove ../springboot-example-svc-hotfix
The Mental Model
The easiest way to think about worktrees is:
Instead of switching the same workspace between branches, give each branch its own workspace.
That means:
One Git repository
│
├── Feature workspace
│ feature/customer-search
│
└── Hotfix workspace
hotfix/invalid-customer-id
Both can be open at the same time.
This becomes even more useful when using coding agents: one workspace can be used by you, while another agent works independently on a hotfix or another feature.

Key Commands
git worktree list git worktree add -b new-branch ../new-folder origin/main git worktree remove ../new-folder
Final Thought
git stash and git switch are still useful.
But when you need to work on two branches in parallel, Git worktrees are often a much cleaner option.
As AI coding agents become part of everyday development, parallel work becomes much more common. Git worktrees give each developer or agent an isolated workspace, so one can work on a feature while another handles a hotfix without stepping on the same files.
One repository. Multiple branches. Separate workspaces. Less context switching.
