A ComfyUI proxy node that acts as a bridge to forward image generation tasks to a remote ComfyUI server. This is an ideal solution for scenarios where local GPU computing power is insufficient and you want to leverage a more powerful remote GPU server seamlessly within your local workflow.
- Place the
ComfyUI-RemoteProxydirectory into your ComfyUI'scustom_nodes/folder. - Install the required Python dependencies:
pip install websocket-client
- Restart ComfyUI.
- Locate the node in the ComfyUI node menu under Remote → Remote ComfyUI Executor.
- Enter the address of your remote ComfyUI server (e.g.,
http://192.168.1.100:8188). - Select the
default_txt2imgtemplate from the dropdown menu. - Fill in the desired parameters such as prompts, width, and height.
- Execute the queue.
- Design your desired workflow over on the remote ComfyUI instance.
- Save it by clicking File → Export (API) to download the workflow JSON.
- In your local ComfyUI, select the
customtemplate on theRemote ComfyUI Executornode. - Paste the exported JSON content into the
custom_workflowtext box. - In your JSON text, use the following placeholder syntax to denote parameters that should be dynamically replaced during execution:
{{positive_prompt}}- Positive prompt{{negative_prompt}}- Negative prompt{{seed}}- Random seed{{width}}/{{height}}- Image resolution dimensions{{input_image}}- The filename of the uploaded input image (for img2img workflows)
You can create your own reusable templates by placing any valid ComfyUI API workflow JSON file into the workflows/ directory. It will automatically be detected and become selectable from the node's template dropdown menu upon restart.
| Parameter | Type | Description |
|---|---|---|
| server_url | STRING | The URL of the remote ComfyUI server. |
| template | COMBO | The workflow template to execute. |
| positive_prompt | STRING | The positive text prompt. |
| negative_prompt | STRING | The negative text prompt. |
| seed | INT | Random seed for generation. |
| width | INT | Image width in pixels. |
| height | INT | Image height in pixels. |
| timeout | INT | Execution timeout in seconds (default: 300). |
| custom_workflow | STRING | Custom workflow JSON (used when template is set to 'custom'). |
| input_images | IMAGE | Input images (Optional, used for img2img pipelines). |
- Ensure the remote ComfyUI server is running and accessible from your local network.
- The
default_txt2imgtemplate defaults to using thev1-5-pruned-emaonly.safetensorsmodel. Please modify the template or use a custom JSON if you need to use a different checkpoint. - During img2img operations, input images are automatically uploaded to the remote server's
inputdirectory prior to execution.