comfyui-dynamic-anynode Introduction
comfyui-dynamic-anynode provides four all-in-one nodes for ComfyUI custom sampling workflows: a guider, sampler, scheduler, and noise node. Each one lets you choose from the compatible implementations available in your ComfyUI setup. When you make a choice, the node shows the inputs that implementation needs.
This helps keep a workflow easier to read. Instead of placing several alternative sampler nodes on the canvas while deciding which one to use, you can work with a single Any Sampler (All-In-One) node and change its selection as needed. The same idea applies to guiders, schedulers, and noise nodes.
Compatibility note: The author notes that you will probably need the ComfyUI
v3-improvementsbranch. The Any nodes detect v3 nodes only, although the author says these include most ComfyUI custom sampling nodes. An installed node that does not appear as a choice may not use the supported v3 format.
How comfyui-dynamic-anynode Works
Think of each Any node as a holder for one part of a custom sampling workflow. For example, Any Sampler is not a single fixed sampler: its selection menu lets you choose a compatible sampler implementation from the nodes in your setup.
The extension inspects the available nodes and uses ComfyUI’s dynamic combo inputs feature to adjust the Any node’s interface. In practice:
- Add an Any node for the part of sampling you want to control.
- Choose an implementation from its combo widget, or selection menu.
- The node updates to show the inputs required by that choice.
- Connect the node into your
CustomSamplingAdvancedworkflow.
The exact choices and inputs you see depend on the compatible custom sampling nodes available in your ComfyUI setup. Changing the selection may therefore change which controls you need to fill in.
comfyui-dynamic-anynode Features
- Any Guider (All-In-One): Lets you select a compatible guider implementation. Use it when you want to choose the guider used in a
CustomSamplingAdvancedworkflow. - Any Sampler (All-In-One): Lets you select a compatible sampler implementation without adding a separate node for every sampler you might try.
- Any Scheduler (All-In-One): Lets you select a compatible scheduler implementation within the same all-in-one interface.
- Any Noise (All-In-One): Lets you select a compatible noise implementation for your custom sampling workflow.
You can add one or more of these nodes and connect them to CustomSamplingAdvanced as appropriate for your workflow. You do not have to use all four.
The main customization is the combo widget on each Any node. Selecting a different implementation updates the node to display that implementation’s required inputs. For example, if you switch the choice in Any Sampler, check the newly displayed inputs before running the workflow; they may differ from those of your previous selection.
Troubleshooting comfyui-dynamic-anynode
An implementation I expected is missing from the selection menu.
- Check whether it is a v3 custom sampling node. The author states that this extension detects v3 nodes only.
- Check that the node is available in your ComfyUI setup.
- If it is available but does not appear, keep the extension’s v3-only limitation in mind; this README does not establish support for older node formats.
The Any nodes or their changing inputs do not work as expected.
- Check whether your ComfyUI setup meets the author’s compatibility note: you will probably need the
v3-improvementsbranch. - Select an implementation using the Any node’s combo widget.
- Look for the inputs required by that selection and connect or fill them in before using the workflow.
Which Any nodes should I add?
Add the ones that match the parts of your CustomSamplingAdvanced workflow you want to select dynamically. You can use one Any node or combine several; the original instructions do not require all four.
Does the extension include its own AI models?
No models are listed in the provided documentation. The extension is described as a way to select among compatible custom sampling nodes in your setup, not as a collection of image-generation models.
