BD Load Audio

ComfyUI has a whole audio API, but it's quiet about it - and getting a WAV into the graph is one of those things you'd expect to be built in and then find isn't obvious. BD_LoadAudio is exactly that missing bit: it loads WAV, MP3, or FLAC from a file path and hands you a standard AUDIO tensor you can wire into anything that consumes audio (save nodes, cache nodes, or audio pipelines).

It's part of the BrainDead pack's Cache/file-ops family - the mirror of BD Cache Audio and BD Save File. The pack ships an audio cache node, so the round trip exists: cache audio out as WAV, then BD_LoadAudio brings it back later without re-running whatever generated it. If you're doing voice work, music loops, or any pipeline where audio and image are generated together, this is the loader you'll reach for.

How it works

One input: file_path (a STRING, absolute path). It loads the file with torchaudio, then returns:

• audio - the standard ComfyUI AUDIO dict: {waveform, sample_rate}, where waveform is shaped (1, channels, samples) so it matches what ComfyUI's audio nodes expect.
• status - a text line that also reports the loaded duration in seconds.

The duration in the status is a small quality-of-life thing: a quick sanity check that the file is what you think it is before it hits downstream.

When it's the right node

• You need a specific audio file on disk in a workflow. Type the path, done.
• Caching. Generate audio once, cache it with the pack's audio cache node, and reload it with this node on later runs to skip regeneration.
• Reproducible pipelines. If your audio comes from a known file (a loop, a VO take), loading it from disk beats regenerating it each run.

The torchaudio dependency is the one real requirement - ComfyUI ships it in most installs, but if you're on a minimal setup you may need pip install torchaudio. The node reports a clear "torchaudio not installed" error if it's missing.

Install

Part of the BrainDead pack:

cd ComfyUI/custom_nodes
git clone https://github.com/BizaNator/ComfyUI-BrainDead
cd ComfyUI-BrainDead
pip install -r requirements.txt

Or search "BrainDead" in ComfyUI Manager and restart.

Troubleshooting

• "File not found" - it's a bare path check; make sure the path exists and is absolute (relative paths don't get ComfyUI-folder treatment here like BD Load Image's do in some contexts).
• "torchaudio not installed" - pip install torchaudio into your ComfyUI environment and restart.
• Unsupported format error - it covers WAV/MP3/FLAC; anything else will fail at load. Convert first.