Sessions broken

### Custom Node Testing

- [x] I have tried disabling custom nodes and the issue persists (see [how to disable custom nodes](https://docs.comfy.org/troubleshooting/custom-node-issues#step-1%3A-test-with-all-custom-nodes-disabled) if you need help)

### Expected Behavior

When I launch my browser (Edge), I usually have a dozen tabs or so from my last session.

Each tab loads its own settings, its own workflow, and I continue where I left off. This has worked since the first version of ComfyUI.

### Actual Behavior

I launch my browser, ALL the tabs are a duplicate of the most recently created tab. Which is a total nightmare.

Even duplicating a tab in the browser produces a duplicate of the most recent tab and not of the tab that was (supposed to be) duplicated.

There is nothing in the logs, so I haven't included any.

### Steps to Reproduce

Launch browser with previous session OR duplicate any ComfyUI tab.

### Debug Logs

```powershell
No logging was created. Here is my bootup log:

call venv/Scripts/activate.bat
Checkpoint files will always be loaded safely.
WARNING: You need pytorch with cu130 or higher to use optimized CUDA operations.
Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
Found comfy_kitchen backend cuda: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
Total VRAM 12288 MB, total RAM 65444 MB
pytorch version: 2.8.0+cu128
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 3060 : cudaMallocAsync
Using async weight offloading with 2 streams
Enabled pinned memory 29449.0
Using pytorch attention
Python version: 3.10.6 (tags/v3.10.6:9c7b4bd, Aug 1 2022, 21:53:49) [MSC v.1932 64 bit (AMD64)]
ComfyUI version: 0.12.3
ComfyUI frontend version: 1.38.13
[Prompt Server] web root: I:\AI\ComfyUI\venv\lib\site-packages\comfyui_frontend_package\static
I:\AI\ComfyUI\venv\lib\site-packages\kornia\feature\lightglue.py:44: FutureWarning: `torch.cuda.amp.custom_fwd(args...)` is deprecated. Please use `torch.amp.custom_fwd(args..., device_type='cuda')` instead.
@torch.cuda.amp.custom_fwd(cast_inputs=torch.float32)
Skipping loading of custom nodes
Context impl SQLiteImpl.
Will assume non-transactional DDL.
Assets scan(roots=['models']) completed in 0.311s (created=0, skipped_existing=1142, orphans_pruned=200, total_seen=1183)
Starting server

To see the GUI go to: http://127.0.0.1:8188
```

### Other

_No response_