I’ve been programming a long, long time. I’m just now beginning to experiment with developing with agentic AI. My first project is a web app I’ve been meaning to develop for a while. So far, I’ve used Claude Opus to generate a nice-looking development plan. I’d like to compare different models for the actual coding from this plan. If I duplicate the project folder and work on the different copies with different models, are the chat sessions isolated from each other? Or will they all be intertwined?
Hey, thanks for the question. You don’t need to duplicate the folder. In Cursor, each chat tab is isolated: its own history, its own context, and its own model selection.
Do this:
- Create 2 chat tabs and pick a different model in each: Tabs | Cursor Docs
- Or duplicate the chat from the needed message and then continue branches with different models: Duplicate | Cursor Docs
Code context is formed from the current workspace, indexing is separate for each open project: Codebase Indexing | Cursor Docs
If it’s convenient, I can suggest how to quickly switch models and run parallel branches.
That would be great, thanks.
Great. The fastest way:
-
Parallel branches:
- Create 2 chat tabs
Cmd+T. In each tab, pick your model in the chat header’s model selector. Switch between tabs withCtrl+Tab: - Or fork a branch from the needed message: three dots → Duplicate Chat:
- Create 2 chat tabs
-
Quick model switch:
- Change the model via the model picker using the shortcut
Cmd+Alt+/. - Keyboard Shortcuts | Cursor Docs
- Change the model via the model picker using the shortcut
-
Parallel run:
- Run both scenarios in separate tabs. For larger parallel runs, check out Cloud Agents: