VRGDG_PromptSplitterV2

VRGDG_PromptSplitterV2 is the plain-text, all-at-once member of the pack's splitter family: multiline prompt text in, sixteen STRING outputs out, no index to fiddle with. It's the closest thing to a straight upgrade of the original VRGDG_PromptSplitter - same idea (chunks of a big prompt onto separate ports), but with the fixed sixteen-output shape the pack's later workflows standardized on.

Where the index-based splitters (ForFL, ForManual, etc.) pull one scene per run, this one dumps the whole batch at once. That makes it the right choice when you want every scene's prompt available simultaneously - typical when you're wiring sixteen parallel generation chains, or when the number of scenes is small enough that you don't need loop logic.

How it works

One required input:

• prompt_text - your multiline prompt block, with each scene's prompt delimited on its own line (the format the pack's LLM node and template builder produce).

Outputs: text_output_1 through text_output_16, all STRING. Wire the ones you need; unused ports just sit empty. If your prompt has fewer than sixteen scenes, the later outputs are simply blank - no error, no drama.

How it differs from the original splitter

The original takes raw text plus a scene_count and lets you dial the number of chunks up to fifty. The V2 drops the count widget and fixes the shape at sixteen, and it's built to consume the pack's later prompt-map text format rather than arbitrary blocks. If you're running a current shipped workflow, use the V2; if you're adapting an older one, keep whatever it references. Neither is "more correct" - they're different generations of the same pattern.

Installing it

With the pack:

cd ComfyUI/custom_nodes
git clone https://github.com/vrgamegirl19/comfyui-vrgamedevgirl

restart, or ComfyUI Manager → search vrgamedev. No models, no special dependencies.

Gotchas

Because there's no index and no count, everything rides on the text format - if your lines aren't delimited the way the node expects, you'll get weird chunk boundaries, and there's no UI to tell you. Also: no validation. Feed it malformed text and you get malformed outputs; run questionable JSON through VRGDG_PromptMapJsonFixer first.