Analyze Audio ♾️Mixlab

I'll be straight with you: AnalyzeAudio is one of the mixlab pack's thinly-documented utility nodes, and its node signature is deliberately generic - it takes a wildcard input and returns a wildcard output. So rather than pretend I can enumerate parameters that don't exist, here's what it actually is and where it fits, grounded in what the node exposes and the pack it lives in.

It sits in the audio corner of shadowcz007's mixlab pack, the same neighborhood as SpeechRecognition and the pack's real-time voice tooling. The pack leans hard into interactive, audio-reactive workflows - voice-driven prompting, live demos, digital-human loops - and AnalyzeAudio is a processing step inside that chain. Its job is to take an audio-related payload (analysis data about a clip - amplitude, timing, the kind of structured audio info the pack's other nodes emit) and pass a processed result on to whatever consumes it next. Think glue inside an audio pipeline, not a headline feature you build a workflow around.

How it works

The * input and * output tell the real story: this node is intentionally type-agnostic so it can slot into a chain wherever the shape of the data suits it. It receives a JSON-shaped audio payload, does its processing, and emits a result of the same open type. Because both ends are wildcards, ComfyUI won't stop you wiring it almost anywhere - which is flexible and also means you are responsible for feeding it something it can actually make sense of.

The inputs and outputs that matter

• json (*) - the input payload, a wildcard type. In practice this is the audio-analysis data produced by an upstream audio node in the pack.
• Output result (*) - the processed result, again wildcard-typed, to hand to the next node in the chain.

There are no beginner knobs here - nothing to set, no thresholds, no mode. It's a plumbing node.

How to install it

Via ComfyUI Manager: search comfyui-mixlab-nodes, install, restart. Or clone the pack:

cd ComfyUI/custom_nodes
git clone https://github.com/shadowcz007/comfyui-mixlab-nodes.git

then pip install -r requirements.txt (or install.bat on the Windows portable build) and restart.

Common issues & troubleshooting

It errors or returns nothing. The wildcard input is the trap: ComfyUI lets you connect anything to a * port, so it's easy to feed it something it wasn't built to parse. Make sure the thing upstream is the pack's own audio-analysis output, not an arbitrary image or string you happened to have a free wire for.

I can't figure out what to wire it to. That's a fair reaction - this is an internal utility, and the pack doesn't document it as a standalone. The honest move is to find the mixlab audio workflow it appears in (the pack ships example audio/voice workflows) and use it in the position that workflow places it, rather than dropping it into a graph cold and guessing at its contract.

Is this the node I actually want? If your goal is transcription, that's SpeechRecognition, not this. AnalyzeAudio is a downstream processing step for audio data, not a speech-to-text or audio-loading node. If you're not already deep in the pack's audio chain, you probably don't need it.