Save Audio (FLAC) (DEPRECATED)

Save Audios (FLAC) is the plural-named sibling of the pack's Save Audio node, and it carries the same banner: DEPRECATED. Same shape, same caveats - it takes an AUDIO tensor, writes it to your output directory as FLAC, and passes the tensor through so the node can live mid-graph. The "s" in the name is mostly about intent (batch contexts), not a different code path: in the source it's the base class that the singular Save Audio subclasses.

If you read the Save Audio article, you already know the drill. This entry exists so the two half-twin nodes are covered, and so the warning is on record twice: these are the old audio savers, and building new work on deprecated nodes is how you end up with a workflow that breaks on the next pack update.

How it works

It subclasses ComfyUI's core SaveAudio, so the mechanism is stock: write the waveform to ComfyUI/output/<filename_prefix>.flac (default prefix audio/ComfyUI), return the tensor unchanged on the audio output. The AUDIO input expects the {waveform, sample_rate} dict, which the pack's Input Audio node produces directly.

Inputs

• audio - the AUDIO tensor to save.
• filename_prefix - path/name, default audio/ComfyUI.

Installing it

Same one-pack install:

cd ComfyUI/custom_nodes
git clone https://github.com/myshell-ai/ComfyUI-ShellAgent-Plugin

restart, or search ComfyUI-ShellAgent-Plugin in ComfyUI Manager. No models; audio support comes from ComfyUI core (torchaudio), not the pack.

Common issues

• Only writes FLAC. There's no format selection - if you need anything else, this isn't it.
• Deprecated. The real risk is future breakage. Don't anchor a workflow you care about to this node.
• Wrong folder? Follow filename_prefix; audio/ComfyUI → output/audio/ComfyUI.flac.

Two deprecated FLAC savers, one recommendation: use them for a quick check if you must, and keep an eye out for whatever the pack ships to replace them.