Perform video editing operations using FFmpeg — cut clips, remove silence, burn captions, generate platform variants. USE THIS SKILL when the user wants to edit video files, extract clips, remove silent parts, add subtitles, or create social media versions. Trigger phrases include "cut video", "extract clip", "remove silence", "burn captions", "add subtitles", "generate variants", "TikTok version", "YouTube Shorts".
This skill provides FFmpeg-based video editing tools for common operations:
| Scenario | Tool |
|---|---|
| Cut a clip from video | extract_clip |
| Find silent parts | detect_silence |
| Remove dead air | remove_silence |
| Add burned-in subtitles | burn_captions |
| Create TikTok/YouTube/Instagram versions | generate_variants |
extract_clipExtract a clip with frame-accurate cutting.
Input:
videoPath (string, required): Source video pathstart (number, required): Start time in secondsend (number, required): End time in secondsoutput (string, required): Output file pathOutput: Confirmation with output path
detect_silenceFind silent regions in video audio.
Input:
videoPath (string, required): Video file paththreshold (string): Silence threshold (default: "-30dB")minDuration (number): Minimum silence duration in seconds (default: 0.5)Output: JSON array of silent regions with start, end, duration
remove_silenceRemove all silent regions from video.
Input:
videoPath (string, required): Video file pathoutput (string, required): Output file paththreshold (string): Silence threshold (default: "-30dB")minDuration (number): Minimum silence duration (default: 0.5)Output: Confirmation with regions removed count and total time saved
burn_captionsHard-code subtitles into video (ASS or SRT format).
Input:
videoPath (string, required): Video file pathcaptionsFile (string, required): Path to ASS/SRT caption fileoutput (string, required): Output file pathOutput: Confirmation with output path
generate_variantsCreate platform-specific aspect ratio variants.
Input:
videoPath (string, required): Video file pathplatforms (array, required): Target platforms — tiktok, youtube, instagram, linkedin, twitteroutputDir (string, required): Output directoryOutput: List of generated variant files
Aspect Ratios:
Extract a clip:
Cut from 10 seconds to 45 seconds of video.mp4
→ extract_clip with start: 10, end: 45
Remove silence:
Clean up the dead air in my recording
→ remove_silence
Create social media versions:
Generate TikTok and YouTube Shorts versions
→ generate_variants with platforms: ["tiktok", "youtube"]