Crop Audio+

Crop Audio+ is the smallest kind of useful: a trim node for an AUDIO signal, with exactly two sliders. You feed it an audio clip, you set the in and out points in seconds, and it hands back the trimmed audio plus its new duration. No browser, no preview player, no model. If your workflow's audio is longer than the video it's meant to accompany - which happens constantly with video models that generate a few seconds of padding - this is the node that fixes the mismatch.

It lives in ComfyUI-FBnodes, François Beaudry's utility pack, and it's a deliberate companion to the pack's video-side machinery: the README frames it as the thing you use to trim audio coming out of Get Video Components+ or an LTX audio decode.

How it works

The node takes the AUDIO's waveform and sample rate, slices between in_point and out_point (both in seconds), and returns a new AUDIO dict plus the trimmed length as a FLOAT. Two details make it friendlier than it looks:

• out_point of 0 means "full length" - so the node's default state does nothing until you actually set an out point, which is the right default for a trim node.
• It re-runs whenever the input audio or the trim points change (it declares its own change signature), so sliding the handles always reflects fresh audio.

duration (FLOAT, seconds) is the genuinely useful output - wire it into a filename template or a display node if you want to know exactly how long your trimmed track is.

The three inputs

• audio (AUDIO) - whatever's feeding the track: decoded LTX audio, a video's extracted audio, a load-audio node.
• in_point (FLOAT, 0+ seconds) - where the clip starts.
• out_point (FLOAT, 0+ seconds) - where it ends; 0 means "don't trim the tail."

Outputs: audio (AUDIO, trimmed) and duration (FLOAT).

Installing it

Same pack, same steps. ComfyUI Manager - search "ComfyUI-FBnodes" - or:

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

Restart ComfyUI. Dependencies are av and color-matcher; nothing heavy, no downloads.

Where it fits and where it bites

The natural spot: between a video's audio extraction and Save Video+, when the generated video runs 5 seconds but the audio track is 6, or vice versa. Pair it with Audio Mono to Stereo on the way out and you've covered the two most common audio cleanup tasks without a single third-party audio pack.

One honest gotcha: it trims, it doesn't resample or normalize. If your audio and video are different sample rates or your track is too quiet, this node won't touch that - it's a pure time-domain slice. And because out_point 0 means "full length," if you genuinely want a zero-length tail you'll be reaching for a value like 0.01, not 0. Minor, but it's the kind of thing that confuses people the first time.