- image
- mask
- latent
- model
- vae
- clip
- positive
- negative
- sdxl tuple
- image
- mask
- latent
- model
- vae
- clip
- positive
- negative
- sdxl tuple
The aegisflow Multi_Pass (displayed as "multi pass") is the "if you only install one passer" node of the AegisFlow Utility pack. Where the single-type passers move one kind of data, this one gangs up almost everything an SD 1.5 workflow touches - image, mask, latent, model, VAE, CLIP, positive and negative conditioning, even an SDXL tuple - onto a single card with matching outputs. In AegisFlow Shima, the multipasser is the part you collapse into the "superlaser": one compact node standing in for a whole cell's worth of connections.
Why this exists
The whole passer family exists because ComfyUI and some common custom nodes try to "fix" open required inputs by auto-wiring whatever matching output is lying around - and sometimes wire the wrong thing, breaking the workflow in ways that are hard to spot. Passers dodge that because every input is optional, so nothing triggers the autofixing. On top of that, they're nameable targets for Use Everywhere / Anything Everywhere style invisible-broadcast nodes: rename this node, point your sender at it, and the whole stack arrives at runtime without a single visible wire.
How it works
Nothing clever. Each of the nine optional inputs passes straight through to the matching output in a fixed order. Leave a socket empty and that output is None. That's the point of a placeholder - it keeps a group's receptor sockets plugged while a modular template is still being built, and it gives the invisible-wiring system one named place to deliver to.
Inputs and outputs
All optional: image (IMAGE), mask (MASK), latent (LATENT), model (MODEL), vae (VAE), clip (CLIP), positive (CONDITIONING), negative (CONDITIONING), sdxl tuple (SDXL_TUPLE). Outputs mirror them one-for-one: image, mask, latent, model, vae, clip, positive, negative, sdxl tuple. There's no processing - this is routing, not rendering.
Installing it
Same as the rest of the pack:
cd ComfyUI/custom_nodes
git clone https://github.com/aegis72/aegisflow_utility_nodes
Restart ComfyUI afterwards. ComfyUI Manager also finds it by searching "AegisFlow Utility Nodes". No models to download; the pack is plain Python on torch/PIL.
Where people get burned
Don't wire an empty output into something that actually consumes it - you'll be feeding None into a KSampler. The multipasser is an anchor for a sender to fill, or a tidy bundle for values you genuinely wire in. If you're on SD 1.5 and want one inbound point for a modular section, this is the one to grab; the XL variant adds the refiner sockets and is the pick for SDXL stacks. Community reports of AegisFlow-era workflows have flagged that ComfyUI Manager doesn't always auto-detect every node in this family, so if a workflow says a node is missing after install, restart once and check the console log before assuming the pack failed.
Inputs (9)
| Name | Type | Default | Description |
|---|---|---|---|
| imageopt | IMAGE | — | |
| maskopt | MASK | — | |
| latentopt | LATENT | — | |
| modelopt | MODEL | — | |
| vaeopt | VAE | — | |
| clipopt | CLIP | — | |
| positiveopt | CONDITIONING | — | |
| negativeopt | CONDITIONING | — | |
| sdxl tupleopt | SDXL_TUPLE | — |
Outputs (9)
| Name | Type | Description |
|---|---|---|
| image | IMAGE | — |
| mask | MASK | — |
| latent | LATENT | — |
| model | MODEL | — |
| vae | VAE | — |
| clip | CLIP | — |
| positive | CONDITIONING | — |
| negative | CONDITIONING | — |
| sdxl tuple | SDXL_TUPLE | — |