Engine Ready

Cut the silence. Keep the story.

Automated jump-cut editing for talking heads, podcasts, and video presentations. Processed entirely on your device.

Drop your video here, or browse filesMP4, WebM, or QuickTime MOV up to 4K
01

Detect

Calculates acoustic loudness and Voice Activity Detection (VAD) to isolate dead air, awkward pauses, and filler hesitations.

02

Review

Instantly preview the cut pacing in real time. Switch between Cleaned and Original audio without waiting for video renders.

03

Export

Local FFmpeg.wasm concatenates kept intervals with PTS/DTS normalization, generating a synchronized H.264/AAC MP4 on your machine.

Interactive Comparison

Before vs. After AutoCut

28.4s saved • 4 cuts
Original Uncut
Speech 1
Silence 1.8s
Speech 2
"um"
Speech 3
AutoCut Result
Speech 1
Speech 2
Speech 3
Technical Specifications & Memory Safety
Audio IngestionDownsampled 16kHz mono Web Audio buffer. Raw PCM is purged immediately after RMS energy calculation to preserve browser heap.
Acoustic Analysis30ms hop RMS energy profiling + dynamic 10th-percentile ambient noise floor + Zero-Crossing Rate VAD gating.
Safety ValidatorEnforces lead-in (80–150ms) and lead-out (100–200ms) consonant padding, 350ms min kept duration, and head/tail anchors.
Video RenderingFFmpeg.wasm 0.12 with explicit PTS/DTS normalization (`setpts=PTS-STARTPTS`, `asetpts=PTS-STARTPTS`) and H.264/AAC encoding.