comfyui-audio-silence-trim

ComfyUI Audio Silence Trim

A small ComfyUI custom node that shortens long silent sections in AUDIO inputs while keeping a configurable portion of each detected silence.

tests

What It Does

• Detects silence from the mean absolute waveform amplitude.
• Keeps part of each silence using keep_ratio.
• Caps retained silence with max_keep_seconds.
• Leaves short cuts untouched when the removed duration is below min_cut_seconds.
• Returns both the trimmed AUDIO object and a JSON report.

Install

Clone this repository into your ComfyUI custom nodes directory:

cd ComfyUI/custom_nodes
git clone https://github.com/gaoqi125/comfyui-audio-silence-trim.git

Restart ComfyUI. The node appears as Audio Silence Trim in the Audio category.

Test

python3 -m unittest discover -s tests -v

The tests use NumPy arrays and do not require a running ComfyUI instance.

Support and Security

Use GitHub Issues for reproducible bugs and Discussions for usage questions. See SECURITY.md and SUPPORT.md before sharing logs or sample media.

Defaults

• keep_ratio: 0.5
• max_keep_seconds: 0.30
• silence_threshold_db: -38.0
• min_silence_seconds: 0.18
• min_cut_seconds: 0.05