ADV Audio Series

The ADV Audio Series (4) node covers the modest case. This is the "I generate eight and I'm going to be picky about it" case. ADV Audio Series takes up to eight AUDIO inputs, bundles whatever's connected into one list, and hands it to the ADV Pick Any Chooser from the same pack - where each clip becomes a row with its own play button.

It's the 8-input sibling in the pmarmotte2/Comfyui_Pick_Any pack. Same mechanism, same audio-dict conventions, just double the room. If your story or sound-design workflow produces a pile of candidate clips per scene, this is the one you reach for instead of its four-slot cousin.

How it works

Identical to ADV Audio Series (4) under the hood. Each slot is an AUDIO input (the standard {waveform, sample_rate} dict), and the node splits any batched waveform tensor into individual clips, drops unconnected None slots, and emits the survivors as one list through the audios output. The chooser converts each entry to WAV on the fly so you can audition them before deciding.

Inputs and outputs

• audio_1 through audio_8 - all optional. Fill as few or as many as your workflow produces; empty slots vanish from the output list.
• audios - the output, a list of audio entries, wired into the chooser's audios port.

Install

One install covers the whole pack:

cd ComfyUI/custom_nodes
git clone https://github.com/pmarmotte2/Comfyui_Pick_Any

Restart ComfyUI and hard-refresh the browser. No requirements.txt, no models, no API keys. Folder name without spaces, as the README asks.

Common issues

• "I connected six and only see four." Count again - a connected-but-empty batch or a None slot gets filtered. Only actual audio content survives the bundling.
• "This is just the 4-node with more slots." Yes, that's exactly what it is, and that's fine. Pick the variant whose slot count matches your workflow; there's no functional difference between them.
• The dialogs take a while with eight clips. Each candidate is encoded to WAV for preview. Keep previews small and you'll be fine.