List Splice Ovum

ListSpliceOvum is the current, non-deprecated version of the pack's splice node: remove a chunk of a list starting at a position, optionally insert new items in its place, and get back both the modified list and the elements you cut. It's JavaScript Array.splice semantics on a Python list, and it's the node to reach for when you need to actually edit a list, not just read from it.

The key word there is edit. Unlike the pack's slice and extend nodes, splice mutates py_list in place. That's not a bug - it's the feature that makes iterative list-processing workflows work - but it's the one thing to remember when you wire it up.

Inputs and output

Straight from the schema:

• py_list - the target list, modified in place. Must be a real Python list.
• start - a STRING widget. Blank → 0. Negative → n + start, clamped to [0, n].
• delete_count - a STRING. Blank → delete to the end (n - start). Negative → 0. Clamped to what remains.
• insert_list - items to insert at start. Blank → nothing inserted. A list → its items. A single non-list value → wrapped as one element.

Outputs:

• * - the same, now-mutated list, so you can thread it forward.
• * - a list of the removed elements, for when you want to process exactly what got cut.

The semantics that will bite you

• Blank delete_count deletes to the end. Leave it empty and you've truncated the list from start onward. Type it deliberately.
• start is clamped, not erroring. Negative values count from the end; out-of-range values clamp. You won't get an out-of-bounds crash, you'll get a silently different splice - verify your math.
• In-place mutation is visible everywhere. If the same underlying list also feeds another branch of the graph, that branch sees the splices too. When you want a copy instead, use ListSliceOvum (non-mutating) or make a copy upstream.

Where it shines

• Consuming lists in loops. Process the first N items, splice them out, pass the shortened list to the next iteration - the "removed" output tells you what you just handled.
• Swapping a section. Remove a bad range of seeds and insert a fresh batch at the same spot in one step.
• Structured queue management. Keep a to-do list in the graph and carve completed work out of it.

Install

Ships in comfy-ovum:

cd ComfyUI/custom_nodes
git clone https://github.com/sfinktah/comfy-ovum

or ComfyUI Manager → comfy-ovum, then restart. No models, no keys.

Gotchas

Non-list py_list raises ValueError: Input must be a Python list. And if you loaded a workflow referencing ListSplice (no suffix), that's the deprecated class - this node is its drop-in replacement with the same inputs and outputs.