Nodes/duckcomfy personal nodes/Conditioning Fallback
ComfyUI Node

Conditioning Fallback

A conditioning that defaults to something sensible

By duckcomfy·Created about a year ago·Updated 10 months ago· 0
Conditioning Fallback
  • primary
  • fallback
  • conditioning

Conditioning Fallback is the "if it's there, use it; if not, use this" node for prompts. It takes an optional primary conditioning and an optional fallback conditioning, and passes primary through when it's connected and non-empty - otherwise you get fallback. No selector, no boolean, no decision on your part. The graph decides for you.

This is the first-non-null fallback pattern, the same shape as rgthree's Any Switch but typed for conditioning. It's the plumbing answer to an annoying reality: a workflow with optional branches produces None when a branch is off, and a sampler fed a None conditioning either errors or silently misbehaves. A fallback catches that. Wire your fancy branch into primary, wire a reliable baseline prompt into fallback, and the sampler always gets a valid conditioning no matter which way the graph resolves.

The inputs and output

  • primary - CONDITIONING, optional. The preferred input.
  • fallback - CONDITIONING, optional. What you get if primary is absent.
  • Output conditioning - primary if present, else fallback.

Both are optional, which is the whole mechanism: an unwired input arrives as None, and the node reads None to mean "not connected." If both are empty, you get a None out and the downstream node will complain - so keep at least fallback wired.

Where it earns its keep

  • Optional enhancement branches. A negative prompt that's only present when some toggle is on - when it's off, the branch emits nothing, and the fallback supplies your standard negative so the sampler never runs unguided.
  • Detailer wiring. A detail pass whose conditioning comes from an expensive prompt branch; when the branch is bypassed, fall back to the main prompt instead of failing.
  • Shared workflows. Someone else's graph may leave the "advanced prompt" input unwired; a fallback means it still generates correctly, which is exactly the resilience you want in a graph you hand to other people.

Install

Part of duckcomfy personal nodes. ComfyUI Manager → search "duckcomfy personal nodes", or:

cd ComfyUI/custom_nodes
git clone https://github.com/duckcomfy/duckcomfy_personal_nodes

Restart ComfyUI. No pip dependencies, no model downloads.

Gotchas

The distinction that bites people: a fallback responds to absence, not to quality. If primary is connected but produces a garbage or empty conditioning, the node passes it through happily - it has no way to judge content. So keep the fallback chain genuinely reliable, and know that a connected-but-blank primary is "present" as far as this node is concerned. Also worth knowing: unlike the pack's CSwitchBoolean* nodes, this one isn't lazy - both branches execute, so don't use it as a performance shortcut around an expensive encode. It's an insurance policy, not a switch.

Categoryduckcomfy

Inputs (2)

NameTypeDefaultDescription
primaryoptCONDITIONING
fallbackoptCONDITIONING

Outputs (1)

NameTypeDescription
conditioningCONDITIONING