Something is wrong with the node "SAM3 Text Segmentation"

[zyd232]

I have noticed that the new version of ComfyUI-SAM3 introduces another plugin of yours, comfy-env, for managing environments.
I encountered some issues.

1. In the “Node Environments” section of the “Environment Manager” panel, the following content is displayed under the “comfyui-sam3” node:

✓ Config: root (W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-sam3\comfy-env-root.toml)
✗ Env: not installed (run comfy-env install)
✓ nodes (W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-sam3\nodes\_env_c034c8)

2. Taking \comfyui-sam3\workflows\image_seg_text.json as an example, it looks like the “SAM3 Text Segmentation” node isn't doing anything - the "mask" output by the node is completely black, and the "visualization" output is the original image. The following is information from the comfyUI backend:

W:\Machine_Learning\ComfyUI_windows_portable>set http_proxy=http://127.0.0.1:8088
W:\Machine_Learning\ComfyUI_windows_portable>set https_proxy=http://127.0.0.1:8088
W:\Machine_Learning\ComfyUI_windows_portable>.\python_embeded\python.exe -s ComfyUI\main.py --windows-standalone-build  --disable-auto-launch --use-sage-attention
[START] Security scan
[DONE] Security scan
## ComfyUI-Manager: installing dependencies done.
** ComfyUI startup time: 2026-04-05 15:18:06.718
** Platform: Windows
** Python version: 3.13.11 (tags/v3.13.11:6278944, Dec  5 2025, 16:26:58) [MSC v.1944 64 bit (AMD64)]
** Python executable: W:\Machine_Learning\ComfyUI_windows_portable\python_embeded\python.exe
** ComfyUI Path: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI
** ComfyUI Base Folder Path: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI
** User directory: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\user
** ComfyUI-Manager config path: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\user\__manager\config.ini
** Log path: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\user\comfyui.log
[comfy-env] comfyui-sam3: 1 isolation env(s):
[comfy-env]   _env_c034c8 -> W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-sam3\nodes\_env_c034c8
[comfy-env] prestartup complete

Prestartup times for custom nodes:
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\rgthree-comfy
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-easy-use
   0.2 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-sam3
   2.9 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager

W:\Machine_Learning\ComfyUI_windows_portable\python_embeded\Lib\site-packages\torch\cuda\__init__.py:65: FutureWarning: The pynvml package is deprecated. Please install nvidia-ml-py instead. If you did not install pynvml directly, please report this to the maintainers of the package that installed pynvml for you.
  import pynvml  # type: ignore[import]
Found comfy_kitchen backend cuda: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_nvfp4']}
Found comfy_kitchen backend triton: {'available': True, 'disabled': True, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8']}
Found comfy_kitchen backend eager: {'available': True, 'disabled': False, 'unavailable_reason': None, 'capabilities': ['apply_rope', 'apply_rope1', 'dequantize_mxfp8', 'dequantize_nvfp4', 'dequantize_per_tensor_fp8', 'quantize_mxfp8', 'quantize_nvfp4', 'quantize_per_tensor_fp8', 'scaled_mm_mxfp8', 'scaled_mm_nvfp4']}
Checkpoint files will always be loaded safely.
Total VRAM 24564 MB, total RAM 65252 MB
pytorch version: 2.10.0+cu130
xformers version: 0.0.35
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 4090 : cudaMallocAsync
Using async weight offloading with 2 streams
Enabled pinned memory 26100.0
working around nvidia conv3d memory bug.
Using sage attention
aimdo: src-win/cuda-detour.c:77:INFO:aimdo_setup_hooks: found driver at 00007FFDB5E70000, installing 4 hooks
aimdo: src-win/cuda-detour.c:61:DEBUG:install_hook_entrys: hooks successfully installed
aimdo: src/control.c:69:INFO:comfy-aimdo inited for GPU: NVIDIA GeForce RTX 4090 (VRAM: 24563 MB)
DynamicVRAM support detected and enabled
Python version: 3.13.11 (tags/v3.13.11:6278944, Dec  5 2025, 16:26:58) [MSC v.1944 64 bit (AMD64)]
ComfyUI version: 0.18.1
comfy-aimdo version: 0.2.12
comfy-kitchen version: 0.2.8
ComfyUI frontend version: 1.42.8
[Prompt Server] web root: W:\Machine_Learning\ComfyUI_windows_portable\python_embeded\Lib\site-packages\comfyui_frontend_package\static
Asset seeder disabled
[Crystools INFO] Crystools version: 1.27.4
[Crystools INFO] Platform release: 11
[Crystools INFO] JETSON: Not detected.
[Crystools INFO] CPU: 13th Gen Intel(R) Core(TM) i9-13900K - Arch: AMD64 - OS: Windows 11
[Crystools INFO] pynvml (NVIDIA) initialized.
[Crystools INFO] GPU/s:
[Crystools INFO] 0) NVIDIA GeForce RTX 4090
[Crystools INFO] NVIDIA Driver: 595.97
[ComfyUI-Easy-Use] server: v1.3.6 Loaded
[ComfyUI-Easy-Use] web root: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-easy-use\web_version/v2 Loaded
ComfyUI-Env-Manager v0.1.0 routes registered
### Loading: ComfyUI-Impact-Pack (V8.28.2)
[Impact Pack] Wildcard total size (0.00 MB) is within cache limit (50.00 MB). Using full cache mode.
[Impact Pack] Wildcards loading done.
### Loading: ComfyUI-Inspire-Pack (V1.23)
### Loading: ComfyUI-Manager (V3.39.2)
[ComfyUI-Manager] network_mode: public
[ComfyUI-Manager] ComfyUI per-queue preview override detected (PR #11261). Manager's preview method feature is disabled. Use ComfyUI's --preview-method CLI option or 'Settings > Execution > Live preview method'.
### ComfyUI Revision: 192 [f21f6b22] *DETACHED | Released on '2026-04-03'
ComfyUI-Multiband loaded: 15 nodes registered
### Loading: ComfyUI-RBG-SmartSeedVariance ###
[comfy-env] Version: 0.2.61
[comfy-env] Removed stale temp dir: C:\Users\zyd232\AppData\Local\Temp\comfyui_pvenv_gizapjly
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/alter-list.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/model-list.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/github-stats.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/extension-node-map.json
[ComfyUI-Manager] default cache updated: https://raw.githubusercontent.com/ltdrdata/ComfyUI-Manager/main/custom-node-list.json
[comfy-env] nodes: sharing host torch from W:\Machine_Learning\ComfyUI_windows_portable\python_embeded\Lib\site-packages
[comfy-env] Importing nodes (isolation root)...
[comfy-env] Cache hit for nodes: 15 nodes
[comfy-env] Registered proxy route: POST /sam3/interactive_segment_one -> nodes.sam3_interactive._api_segment_one
[comfy-env] Imported nodes root: 15 nodes (isolation)
[comfy-env] Registered 15 total nodes
[W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using ckpts path: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux\ckpts
[W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using symlinks: False
[W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux] | INFO -> Using ort providers: ['CUDAExecutionProvider', 'DirectMLExecutionProvider', 'OpenVINOExecutionProvider', 'ROCMExecutionProvider', 'CPUExecutionProvider', 'CoreMLExecutionProvider']
W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux\node_wrappers\dwpose.py:26: UserWarning: DWPose: Onnxruntime not found or doesn't come with acceleration providers, switch to OpenCV with CPU device. DWPose might run very slowly
  warnings.warn("DWPose: Onnxruntime not found or doesn't come with acceleration providers, switch to OpenCV with CPU device. DWPose might run very slowly")
USDU batch patches applied successfully.
USDU batch patches applied successfully.

[rgthree-comfy] Loaded 48 magnificent nodes. 🎉

Total VRAM 24564 MB, total RAM 65252 MB
pytorch version: 2.10.0+cu130
xformers version: 0.0.35
Set vram state to: NORMAL_VRAM
Device: cuda:0 NVIDIA GeForce RTX 4090 : cudaMallocAsync
Using async weight offloading with 2 streams
Enabled pinned memory 26100.0
### Loading: SDPose OOD Nodes ###
⚡ SeedVR2 optimizations check: SageAttention ✅ | Flash Attention ✅ | Triton ✅
🔧 Conv3d workaround active: PyTorch 2.10.0, cuDNN 91200 (fixing VAE 3x memory bug)
📊 Initial CUDA memory: 22.44GB free / 23.99GB total
   🔧 torchaudio.save/load globally patched (scipy for WAV files, no TorchCodec)
   🔇 Suppressed torchaudio 2.9 migration warnings
ℹ️ Critical package versions: NumPy 2.2.6, Librosa 0.11.0, Numba 0.65.0, PyTorch 2.10.0+cu130, TorchAudio 2.11.0a0+d20260219.cu130torch2.11.0cxx11abi1, Transformers 4.57.3, Accelerate 1.13.0, SoundFile 0.13.1

Import times for custom nodes:
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\websocket_image_save.py
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Olm-DragCrop
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_memory_cleanup
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\Comfyui-Resolution-Master
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\sdmatte
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-seamless-tiling
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\cg-use-everywhere
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Env-Manager
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\editutils
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-zyd232-nodes
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\qweneditutils
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_pixel_snapping
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\erosdiffusion-eulerflowmatchingdiscretescheduler
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-RBG-SmartSeedVariance
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\LanPaint
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-enricos-nodes
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyMath
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_essentials
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-custom-scripts
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_ultimatesdupscale
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-segment-anything-2
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-advanced-controlnet
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\rgthree-comfy
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-kjnodes
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Multiband
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI_LayerStyle_Advance
   0.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-inspire-pack
   0.1 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-LBMWrapper
   0.1 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Impact-Pack
   0.1 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\seedvr2_videoupscaler
   0.1 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-advancedliveportrait
   0.1 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_layerstyle
   0.1 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-VideoHelperSuite
   0.2 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Manager
   0.2 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-WanVideoWrapper
   0.2 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui_controlnet_aux
   0.3 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\tts_audio_suite
   0.5 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-Crystools
   0.5 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\ComfyUI-GIMM-VFI
   0.6 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\sdpose-ood
   3.0 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-sam3
   3.3 seconds: W:\Machine_Learning\ComfyUI_windows_portable\ComfyUI\custom_nodes\comfyui-easy-use

Context impl SQLiteImpl.
Will assume non-transactional DDL.
Starting server

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

got prompt
[comfyui-sam3] python: W:\ce\_env_c034c8\.pixi\envs\default\python.exe
[comfyui-sam3] sys_path sent to worker: ['W:\\Machine_Learning\\ComfyUI_windows_portable\\ComfyUI', 'W:\\Machine_Learning\\ComfyUI_windows_portable\\ComfyUI\\custom_nodes\\comfyui-sam3', 'W:\\Machine_Learning\\ComfyUI_windows_portable\\python_embeded\\Lib\\site-packages', 'W:\\ce\\_env_c034c8\\.pixi\\envs\\default\\Lib\\site-packages', 'W:\\Machine_Learning\\ComfyUI_windows_portable\\ComfyUI\\custom_nodes\\comfyui-sam3']
FETCH ComfyRegistry Data: 30/136
[comfyui-sam3] Health check: ping (timeout=5.0s)...
[comfyui-sam3] Health check: ok
FETCH ComfyRegistry Data: 35/136
[worker:comfyui-sam3] WARNING: SageAttention is not compatible with SAM3 (relative position bias modifies Q/K dimensions). Using attention_pytorch instead.
Prompt executed in 10.15 seconds

[zyd232]

I tested the node "SAM3 Point Segmentation", it works well.
But the node "SAM3 Text Segmentation" always output totally black mask and the "visualization" output is the original image. It seems that no matter what I enter in "text_prompt", it doesn't work.

[PozzettiAndrea]

Hi @zyd232 ! What GPU/OS are you working with? :)

[sho9029]

I have the same issue.

• GPU: NVIDIA GeForce RTX 4080
• OS: Windows 11

@zyd232's log also shows RTX 4090 / Windows 11.

[PozzettiAndrea]

@zyd232 @sho9029 Please try deleting your entire ComfyUI-SAM3 folder in custom_nodes, then delete the already existing environment in C:\ce folder, then reinstall from scratch from the ComfyUI Manager. If it still doesn't work, please send me both the installation and the run logs. You can write to me on Discord if it's easier for you guys @AndrewWellies

[zyd232]

@zyd232 @sho9029 Please try deleting your entire ComfyUI-SAM3 folder in custom_nodes, then delete the already existing environment in C:\ce folder, then reinstall from scratch from the ComfyUI Manager. If it still doesn't work, please send me both the installation and the run logs. You can write to me on Discord if it's easier for you guys @AndrewWellies

GPU: NVIDIA GeForce RTX 4090
OS: Windows 11

I deleted the ComfyUI-SAM3 folder and ce folder, reinstalled SAM3 from scratch through the ComfyUI Manager. But the node "SAM3 Text Segmentation" still doesn't work.

As the screenshot shows, the nodes like "SAM3 Point Segmentation" works well, but "SAM3 Text Segmentation" always outputs empty results:

Here is the installation and running log:
ComfyUI.txt

[illusive5943]

Same problem on Linux, SAM3 Text Segmentation passthroughs the source image as-is / unchanged and mask comes out black.