Separating the Agent Manager from the IDE Destroys the Entire Workflow

The only reason I use Antigravity IDE is because the AI agent works directly inside the editor with the split-window workflow. That setup is what makes development fast and practical.

Moving toward a standalone chat app and planning to remove the agent from the IDE in the future feels like a huge downgrade. It breaks local context, interrupts workflow, and makes the whole experience worse for developers.

On top of that, the newer model outputs feel noticeably worse compared to before. If the long-term direction is to permanently separate the agent from the IDE, then I will simply stop using it.

100% That’s exactly why I moved from VScode to Antigravity. This new change is ruining the whole idea.

I should be going back to VSCode.