PyList Item Cast

So you've got a PyList and its items are all the wrong type. Maybe they came out of Split String as text and your next node wants numbers. Maybe you need to flip "0"/"1" strings into real booleans for a toggle. PyList Item Cast is the per-item version of Python's int(), float(), str() and friends - one widget, and every element gets converted.

It's a small node and it does exactly one thing, which is honestly why you'll reach for it: in a language as casually typed as Python, "list of strings" and "list of ints" look identical until a downstream node chokes on the difference.

How it works

Under the hood it's a one-liner: [try_cast(x, TYPE) for x in PYLIST]. Every item gets run through the same conversion and the result comes back as a new PyList. Nothing in-place, nothing fancy.

The two inputs that matter

• PYLIST - the list to convert. forceInput, so it must come from an upstream PyList-producing node; there's no widget.
• TYPE - the target type, chosen from STRING, INT, FLOAT, NUMBER, BOOLEAN, PYLIST.

The output is a single PYLIST socket holding the converted list.

Three things worth knowing about the TYPE choices, because two of them are traps:

• NUMBER is a lie. Per the author's own tooltip, NUMBER is the same as FLOAT - it calls float() and that's it. No "pick whatever fits" smartness.
• PYLIST as a target is a no-op. try_cast has no case for it, so every item passes through unchanged. Selecting it just relabels the socket; it's there for re-typing, not conversion.
• BOOLEAN has opinions. Numbers become True if greater than zero, else False. Strings get parsed as floats first, then compared - so "0.5" becomes True and "" becomes False. Good enough for toggles, just don't assume it matches Python's bool() on every input.

Where the edge cases bite

The big one: INT runs int() on each element, and int("3.7") raises a ValueError, which will kill your queue mid-run. If your strings might carry decimals, cast to FLOAT first - then to INT if you really need integers. Strings that aren't numeric at all will crash either way, and there's no try/except here to save you. That's not a bug in the node; that's Python being Python, but it's the failure mode people actually hit.

Also remember what this node doesn't do: the result is still a PyList, i.e. one value on one socket. It does not start per-item processing. If you need the converted items to feed downstream nodes one at a time, chain PyList To List after this and you're done.

Installing it

Part of godmt/ComfyUI-List-Utils. In ComfyUI Manager, search "ComfyUI-List-Utils" and install, or:

cd ComfyUI/custom_nodes
git clone https://github.com/godmt/ComfyUI-List-Utils

Restart ComfyUI and you'll find it under list_utils. No models, no extra pip dependencies - it's pure Python with no imports beyond the standard library, so there's nothing to download and nothing to break.

In practice

The workflow that'll actually use this: text in, Split String, cast the PYLIST output to INT or FLOAT, then run it through a numeric node that would have refused the string version. It's a utility node, not a showstopper - but when your seeds arrive as "42" and something downstream demands 42, it's the difference between a clean run and a head-scratching afternoon.