EvaluateMultiple6

Same evaluate-your-own-Python mechanism as the rest of this family, sized up to six lanes: in1 through in6 in, out1 through out6 out, one shared expression box running against all of them.

Where six lanes actually earns the extra size

Every EvaluateMultiple variant shares the same input_count widget (1–20), so in principle you could start from EvaluateMultiple1 and drag it up to six every time. In practice, once you're wiring up six separate inputs, having them present from the start is genuinely easier to build with than resizing mid-graph and rewiring sockets that moved. Six is also a size where the use case tends to shift - you're less often doing "one quick calculation" and more often processing a small batch of related values together: six crop-region sizes, six per-tile offsets, six differently-named outputs that all come from the same underlying computation.

Inputs and outputs

• expression (multiline STRING) - blank by default, your Python goes here.
• input_count (INT, 1–20, default 6).
• print_to_console - debug echo to the ComfyUI terminal.
• Optional: in1–in6 (type *).
• Outputs: out1–out6.

Same caution, worth repeating at this size especially

Real Python, no sandbox - the README states plainly that this family runs "any fragment of Python code," in contrast to rgthree-comfy's Power Puter, which deliberately restricts itself to an AST-parsed, allowlisted subset with no file or network access. At six lanes you're also more likely to be writing something with real structure - loops, conditionals, a few intermediate variables. That's fine, it's what the node is for, but it's also the point where a wall of Python jammed into one text box starts costing you readability. A comment at the top of the expression explaining what each inN/outN represents will save future-you real time.

Installing it

ComfyUI Manager: search RuiquNodes, install, restart. Manual:

cd ComfyUI/custom_nodes
git clone https://github.com/ruiqutech/ComfyUI-RuiquNodes

then restart. No models, no extra dependencies.

Common issues

Some outputs are always None. Your expression isn't assigning to every outN you've wired downstream - check each one explicitly rather than assuming an unused lane is harmless; it'll just pass None along silently.

Hard to tell which input is which once wired. With six sockets on one node, label your links or add a note - the node itself only gives you in1...in6, it won't remind you what you meant each one to be three weeks from now.

A crash mid-expression. Same as any Python error: the traceback names the line. Read it first before assuming the node itself is broken.

Only using two or three of the six lanes. That's fine - every inN/outN pair is independent and optional. Grabbing EvaluateMultiple6 for a task that only needs two outputs today, because you expect to grow it later, is a reasonable call; just don't feel obligated to wire up sockets you have no use for yet.