Select Audio From List

Some nodes do one thing and do it so plainly there's almost nothing to explain. This is one of those. Audio Segmenter hands you a list of clips split on silence; Select Audio From List reaches in, grabs the clip at the index you give it, and hands it back as a single AUDIO object. That's the entire job. Sometimes that's exactly the node you need - a splitter produces five clips, and you want clip three wired to a preview or a save node without building a loop contraption.

Inputs and output

Two inputs, one output:

• audio_list - the list from Audio Segmenter (typed *, so ComfyUI won't complain if you mis-wire it).
• index - which clip, 0-based, default 0.
• AUDIO - the selected clip as a standard single-audio object.

One behavior worth internalizing: if the list is empty or the index is out of range, it doesn't throw - it returns a silent one-sample audio placeholder. That's forgiving for lazy graphs, but it also means a bad index fails quietly. If you get blank output, check the index against the count output of the segmenter before you go hunting for deeper problems.

When you'd actually use it

Honestly, for most people the value shows up in one of two patterns. Either you're inspecting individual clips - select one, wire it to an audio preview or save node, confirm the split quality, adjust silence_thresh_db, repeat - or you're building a batch pass where you intentionally iterate one clip at a time. In both cases the node is a dumb, reliable indexer, and that's a compliment.

It pairs with the pack's silence splitter. If you're working with the Whisper-based Audio Segment By Paragraph node instead, you want its sibling Select AudioClip From List, which indexes the AudioClip list and returns both the audio and the paragraph's text - this node gives you audio only. Same idea, different half of the pack; remember that both list inputs are typed *, so it's easy to wire the wrong list into the wrong selector and get silence back.

Installing the pack

ComfyUI Manager: search "Audio Segment" (pack title ComfyUI Audio Segment Plugin). Manual:

cd path/to/ComfyUI/custom_nodes/
git clone https://github.com/ahkimkoo/Comfyui-AudioSegment
cd Comfyui-AudioSegment
pip install -r requirements.txt

Restart ComfyUI and the nodes appear under Audio/Segmentation. Fair warning: the shared requirements.txt is heavy - torch, openai-whisper, transformers, accelerate, pydub, opencc-python-reimplemented - even though this little selector needs almost none of it. You're effectively installing the whole pack to use a two-input indexer. That's the cost of one-install packs with a flagship node; if you only need the silence splitter and its selector, it's still the simplest way to get them, and ComfyUI already ships torch so pip usually just confirms what's there.