Multi-Clip Text Script: Main Simple

Sometimes a node ships because the author wanted exactly one thing. Multi-Clip Text Script: Main Simple is that thing. It's the Main node with prefix, suffix, and negative all stripped out - one input, one parse, two outputs. The README is refreshingly blunt about why: the author didn't really use the extra fields, and mainly used this version to copy a whole script into ChatGPT and ask where it was failing.

What it does. Same script format, same parsing as the full Main node - label sections (clip01), (clip02), … and it regex-parses them, sorts numerically ((clip1), (clip02), (clip10) → 1, 2, 10), and hands you the clips. Internally it stores empty prefix and suffix, so its pair_data encoding is byte-compatible with Main's: pair_data from either node feeds the same Clip Selector. You can swap nodes in a workflow without rewiring anything downstream.

The one input that matters: multi_clip_script. That's genuinely it. There are no optional inputs, no hidden fields, no settings.

Outputs. pair_data (feed it to the Clip Selector) and num_clips (an INT telling you how many shots were found - the cheap sanity check that your labels parsed at all).

Reach for it when:

• You don't need shared boilerplate - you're pasting the script into an LLM anyway, or each shot's prompt already carries its own full description.
• You want the smallest possible surface area: one widget, nothing to accidentally leave filled in.
• You're just testing whether the multi-clip workflow works for you before committing to prefix/suffix structure.

One registration gotcha worth knowing. This node is a v1.1.0 addition (January 2026), and it shipped with a classic amateur-publisher mistake: the author pushed an old __init__.py to GitHub that didn't register the new class, so the node existed in code but never appeared in the menu. Fixed in v1.1.1, but if you cloned the pack in that window it'll look like the node doesn't exist. git pull and restart before you go hunting for a phantom. Current clones are fine - all three classes are registered.

Install is the same pack either way:

cd ComfyUI/custom_nodes
git clone https://github.com/TuonoMindCode/ComfyUI-MultiClip-Text-Script

Restart ComfyUI, search "Multi-Clip Text Script: Main Simple". No models, no dependencies beyond Python's standard library.

Honest take: if prefix/suffix consistency is the whole point of this pack for you, this node misses it - use Main. But if you author prompts in an LLM and just want clean per-shot text out, this is the one I'd reach for. One widget, no decorations, and it makes the "write the script, let the model pick the shot" loop about as frictionless as ComfyUI gets.