📦 Egregora Static Payload Pack

Some custom nodes exist to solve a real everyday problem; others exist to make an architecture complete. Egregora Static Payload Pack is firmly in the second camp. It's one of the three runtime-payload nodes in the Egregora-ARMD pack, and its job is tiny: wrap two static string values into a runtime_payload_adapter that Egregora Adaptive Diffusion Apply can consume.

What it actually does

The ARMD denoising wrapper accepts an optional runtime_payload_adapter, and anything you pack into it gets injected into every regional denoising pass (after being validated against key collisions). Static payloads are the non-spatial half of that story - the companion to Egregora Spatial Tensor Pack, which carries images and latents. This node carries plain key/value string flags.

The interface is four fields:

• name_value_a, name_value_b - the keys (defaults runtime_flag_a / runtime_flag_b).
• value_a, value_b - the values.

One behavior worth knowing: a value left empty is skipped entirely - the payload only includes entries where the value isn't "". So you can leave value_b blank and the adapter simply won't carry that key. There are no other knobs; it's about as bare-bones as a node gets.

When would you actually use this?

Honestly: rarely, unless you're building on top of the pack. The realistic scenario is that you've written (or are following) an ARMD workflow that reads a runtime flag - something like a mode switch or a per-run parameter that the regional mixer's payload hooks look for. You'd pack your flag here, merge it with any spatial tensors using Egregora Runtime Adapter Merge, and hand the result to Adaptive Diffusion Apply. If you're just running a stock upscaling or from-scratch workflow from the pack's README, you can skip this node entirely - the runtime_payload_adapter input on Adaptive Diffusion Apply is optional, and the default paths don't need it.

Installing

It ships with the rest of the pack:

cd ComfyUI/custom_nodes
git clone https://github.com/lucasgattas/ComfyUI-Egregora-ARMD

restart, or install "ComfyUI-Egregora-ARMD" via ComfyUI Manager. No model downloads, dependencies are numpy, opencv-python, PyWavelets and torch. This is the pack's author Lucas Gattas (mrgattax) keeping the plumbing generic rather than hard-coding one workflow - which is good engineering, even if it means a node like this looks pointless until the moment you need exactly it.