Make List | akatz-loops

Some workflows need to treat a handful of separate values as one unit - ten prompts to try, five seed values, a few images you want processed together. ComfyUI has no native "these are a list" button, so you reach for a list-builder node, and MakeListNode is the pack's plainest one: up to ten value inputs in, one list out. value1 is required; value2 through value9 are optional. Connect whatever you have, and the output is a list of just the connected inputs, in order, with no placeholder slots for the ones you left empty.

That last detail is worth knowing: the node builds the list from what's actually connected, skipping unwired sockets. Wire value1, value3, and value7 and you get a three-item list, not a seven-item one with gaps. For most uses that's exactly right - you want a dense list, not a sparse one with Nones to trip up downstream logic.

What it's for

The list output plugs straight into the rest of the pack's list machinery and beyond. It's the natural front end for GetIntFromList / GetFloatFromList when you want to pick one element out by index. It's the input side of the pack's ListToAccumulation if you want to run a loop over a fixed set of values. And because the output is typed as a list of * (any type), the same node can bundle images, ints, strings, or a mix - although a mixed-type list is usually a sign you're about to confuse a downstream node that expects one type.

The most common beginner win is simpler than all of that: you have N values scattered across the graph and you want to select one of them by an index you compute at runtime. Bundle them here, then let an index pick from the list instead of hand-wiring an ever-growing switch node. It's the "hold these together" primitive that makes the pack's list-based selection nodes useful at all.

Install

Part of Akatz-Loop-Nodes (akatz-ai). Via ComfyUI Manager, search "Akatz-Loop-Nodes"; or:

cd ComfyUI/custom_nodes
git clone https://github.com/akatz-ai/Akatz-Loop-Nodes

Restart ComfyUI. The pack's only dependency is opencv-python (used by its image nodes, not this one); no model downloads. The custom_nodes.ComfyUI-Execution-Inversion module path on listings is lineage from BadCafeCode's execution-inversion demo - not a separate package.

Troubleshooting is mostly about expectation: if a downstream node says it got fewer items than you expected, you probably left a value socket unwired and the node correctly skipped it. And remember the output is a list, not a batch - a list of images is different from a batched IMAGE tensor, and some image nodes expect one or the other. When in doubt, check what the consuming node's input type actually is.