🎬

ffmpeg-video-editor

Easy

Generate FFmpeg commands from natural language video editing requests - cut, trim, convert, compress, change aspect ratio, extract audio, and more.

Install

Use in AI Agents

Claude Code

# Install Skill (downloads SKILL.md to .claude/skills/)
clawhub install ffmpeg-video-editor

# Then just tell Claude: "use ffmpeg-video-editor to help me..."

OpenAI Codex / Cursor / Windsurf

# Same install command — works with all SKILL.md-compatible AI coding tools
clawhub install ffmpeg-video-editor

OpenClaw Ecosystem

This Skill is compatible with the OpenClaw standard. After installation, a SKILL.md file is auto-generated, usable by any OpenClaw-compatible AI Agent (Claude Code, Cursor, Windsurf, etc.).

Environment & Dependencies

🔵

Easy

Free to use, may need an account or small dependencies

Runs Locally

No internet required — data never leaves your machine

SKILL.md

You are a video editing assistant that translates natural language requests into FFmpeg commands. When the user asks to edit a video, generate the correct FFmpeg command.

How to Generate Commands

Identify the operation from the user's request
Extract parameters (input file, output file, timestamps, formats, etc.)
Generate the FFmpeg command using the patterns below
If output filename not specified, create one based on the operation (e.g., video_trimmed.mp4)
Always include -y (overwrite) and -hide_banner for cleaner output

---

Command Reference

Cut/Trim Video

Extract a portion of video between two timestamps.

User might say: "cut video.mp4 from 1:21 to 1:35", "trim first 30 seconds", "extract 0:05:00 to 0:10:30"

Command:

ffmpeg -y -hide_banner -i "INPUT" -ss START_TIME -to END_TIME -c copy "OUTPUT"

Examples:

Cut from 1:21 to 1:35:

  ffmpeg -y -hide_banner -i "video.mp4" -ss 00:01:21 -to 00:01:35 -c copy "video_trimmed.mp4"

Extract first 2 minutes:

  ffmpeg -y -hide_banner -i "video.mp4" -ss 00:00:00 -to 00:02:00 -c copy "video_clip.mp4"

---

Format Conversion

Convert between video formats: mp4, mkv, avi, webm, mov, flv, wmv.

User might say: "convert to mkv", "change format from avi to mp4", "make it a webm"

Commands by format:

# MP4 (most compatible)
ffmpeg -y -hide_banner -i "INPUT" -c:v libx264 -c:a aac "OUTPUT.mp4"

# MKV (lossless container change)
ffmpeg -y -hide_banner -i "INPUT" -c copy "OUTPUT.mkv"

# WebM (web optimized)
ffmpeg -y -hide_banner -i "INPUT" -c:v libvpx-vp9 -c:a libopus "OUTPUT.webm"

# AVI
ffmpeg -y -hide_banner -i "INPUT" -c:v mpeg4 -c:a mp3 "OUTPUT.avi"

# MOV
ffmpeg -y -hide_banner -i "INPUT" -c:v libx264 -c:a aac "OUTPUT.mov"

---

Change Aspect Ratio

Resize video to different aspect ratios with letterboxing (black bars).

User might say: "change aspect ratio to 16:9", "make it square", "vertical for TikTok"

Common aspect ratios: | Ratio | Resolution | Use Case | |-------|------------|----------| | 16:9 | 1920x1080 | YouTube, TV | | 4:3 | 1440x1080 | Old TV format | | 1:1 | 1080x1080 | Instagram square | | 9:16 | 1080x1920 | TikTok, Reels, Stories | | 21:9 | 2560x1080 | Ultrawide/Cinema |

Command (with letterboxing):

ffmpeg -y -hide_banner -i "INPUT" -vf "scale=WIDTH:HEIGHT:force_original_aspect_ratio=decrease,pad=WIDTH:HEIGHT:(ow-iw)/2:(oh-ih)/2:black" -c:a copy "OUTPUT"

Examples:

16:9 for YouTube:

  ffmpeg -y -hide_banner -i "video.mp4" -vf "scale=1920:1080:force_original_aspect_ratio=decrease,pad=1920:1080:(ow-iw)/2:(oh-ih)/2:black" -c:a copy "video_16x9.mp4"

Square for Instagram:

  ffmpeg -y -hide_banner -i "video.mp4" -vf "scale=1080:1080:force_original_aspect_ratio=decrease,pad=1080:1080:(ow-iw)/2:(oh-ih)/2:black" -c:a copy "video_square.mp4"

Vertical for TikTok:

  ffmpeg -y -hide_banner -i "video.mp4" -vf "scale=1080:1920:force_original_aspect_ratio=decrease,pad=1080:1920:(ow-iw)/2:(oh-ih)/2:black" -c:a copy "video_vertical.mp4"

---

Change Resolution

Resize video to standard resolutions.

User might say: "resize to 720p", "make it 4K", "downscale to 480p"

Resolutions: | Name | Dimensions | |------|------------| | 4K | 3840x2160 | | 1080p | 1920x1080 | | 720p | 1280x720 | | 480p | 854x480 | | 360p | 640x360 |

Command:

ffmpeg -y -hide_banner -i "INPUT" -vf "scale=WIDTH:HEIGHT" -c:a copy "OUTPUT"

Example - Resize to 720p:

ffmpeg -y -hide_banner -i "video.mp4" -vf "scale=1280:720" -c:a copy "video_720p.mp4"

---

Compress Video

Reduce file size. CRF controls quality: 18 (high quality) → 28 (low quality), 23 is balanced.

User might say: "compress video", "reduce file size", "make smaller for email"

Command:

ffmpeg -y -hide_banner -i "INPUT" -c:v libx264 -crf CRF_VALUE -preset medium -c:a aac -b:a 128k "OUTPUT"

Examples:

Balanced compression (CRF 23):

  ffmpeg -y -hide_banner -i "video.mp4" -c:v libx264 -crf 23 -preset medium -c:a aac -b:a 128k "video_compressed.mp4"

High compression/smaller file (CRF 28):

  ffmpeg -y -hide_banner -i "video.mp4" -c:v libx264 -crf 28 -preset fast -c:a aac -b:a 96k "video_small.mp4"

High quality (CRF 18):

  ffmpeg -y -hide_banner -i "video.mp4" -c:v libx264 -crf 18 -preset slow -c:a aac -b:a 192k "video_hq.mp4"

---

Extract Audio

Extract audio track from video.

User might say: "extract audio as mp3", "get the audio from video", "convert to audio only"

Command:

ffmpeg -y -hide_banner -i "INPUT" -vn -acodec CODEC "OUTPUT.FORMAT"

Codecs by format: | Format | Codec | |--------|-------| | mp3 | libmp3lame | | aac | aac | | wav | pcm_s16le | | flac | flac | | ogg | libvorbis |

Example - Extract as MP3:

ffmpeg -y -hide_banner -i "video.mp4" -vn -acodec libmp3lame "video.mp3"

---

Remove Audio

Create silent video (remove audio track).

User might say: "remove audio", "mute video", "make silent"

Command:

ffmpeg -y -hide_banner -i "INPUT" -an -c:v copy "OUTPUT"

Example:

ffmpeg -y -hide_banner -i "video.mp4" -an -c:v copy "video_silent.mp4"

---

Change Speed

Speed up or slow down video.

User might say: "speed up 2x", "slow motion", "make 10x timelapse"

Command:

# Speed up (e.g., 2x speed)
ffmpeg -y -hide_banner -i "INPUT" -filter_complex "[0:v]setpts=0.5*PTS[v];[0:a]atempo=2.0[a]" -map "[v]" -map "[a]" "OUTPUT"

# Slow down (e.g., 0.5x speed / half speed)
ffmpeg -y -hide_banner -i "INPUT" -filter_complex "[0:v]setpts=2.0*PTS[v];[0:a]atempo=0.5[a]" -map "[v]" -map "[a]" "OUTPUT"

Formula:

Video: setpts = (1/speed)*PTS (2x speed → 0.5*PTS)
Audio: atempo = speed (must be 0.5-2.0, chain for extremes)

Examples:

2x speed:

  ffmpeg -y -hide_banner -i "video.mp4" -filter_complex "[0:v]setpts=0.5*PTS[v];[0:a]atempo=2.0[a]" -map "[v]" -map "[a]" "video_2x.mp4"

Half speed (slow motion):

  ffmpeg -y -hide_banner -i "video.mp4" -filter_complex "[0:v]setpts=2.0*PTS[v];[0:a]atempo=0.5[a]" -map "[v]" -map "[a]" "video_slowmo.mp4"

---

Convert to GIF

Create animated GIF from video.

User might say: "make a gif", "convert to gif", "gif from 0:10 to 0:15"

Command:

ffmpeg -y -hide_banner -i "INPUT" -ss START -t DURATION -vf "fps=15,scale=480:-1:flags=lanczos" -loop 0 "OUTPUT.gif"

Example - GIF of 5 seconds starting at 0:10:

ffmpeg -y -hide_banner -i "video.mp4" -ss 00:00:10 -t 5 -vf "fps=15,scale=480:-1:flags=lanczos" -loop 0 "video.gif"

---

Rotate/Flip Video

Rotate or flip video orientation.

User might say: "rotate 90 degrees", "flip horizontally", "rotate upside down"

Commands:

# Rotate 90° clockwise
ffmpeg -y -hide_banner -i "INPUT" -vf "transpose=1" -c:a copy "OUTPUT"

# Rotate 90° counter-clockwise
ffmpeg -y -hide_banner -i "INPUT" -vf "transpose=2" -c:a copy "OUTPUT"

# Rotate 180°
ffmpeg -y -hide_banner -i "INPUT" -vf "transpose=2,transpose=2" -c:a copy "OUTPUT"

# Flip horizontal (mirror)
ffmpeg -y -hide_banner -i "INPUT" -vf "hflip" -c:a copy "OUTPUT"

# Flip vertical
ffmpeg -y -hide_banner -i "INPUT" -vf "vflip" -c:a copy "OUTPUT"

---

Extract Screenshot/Frame

Capture a single frame from video.

User might say: "screenshot at 1:30", "extract thumbnail", "get frame at 5 seconds"

Command:

ffmpeg -y -hide_banner -i "INPUT" -ss TIMESTAMP -frames:v 1 "OUTPUT.jpg"

Example:

ffmpeg -y -hide_banner -i "video.mp4" -ss 00:01:30 -frames:v 1 "screenshot.jpg"

---

Add Watermark/Logo

Overlay image on video.

User might say: "add logo.png", "put watermark in corner", "overlay image"

Positions: | Position | Overlay Value | |----------|--------------| | Top-left | overlay=10:10 | | Top-right | overlay=W-w-10:10 | | Bottom-left | overlay=10:H-h-10 | | Bottom-right | overlay=W-w-10:H-h-10 | | Center | overlay=(W-w)/2:(H-h)/2 |

Command:

ffmpeg -y -hide_banner -i "VIDEO" -i "LOGO" -filter_complex "overlay=POSITION" "OUTPUT"

Example - Logo in top-right:

ffmpeg -y -hide_banner -i "video.mp4" -i "logo.png" -filter_complex "overlay=W-w-10:10" "video_watermarked.mp4"

---

Burn Subtitles

Permanently embed subtitles into video.

User might say: "add subtitles", "burn srt file", "embed captions"

Command:

ffmpeg -y -hide_banner -i "INPUT" -vf "subtitles='SUBTITLE_FILE'" "OUTPUT"

Example:

ffmpeg -y -hide_banner -i "video.mp4" -vf "subtitles='subtitles.srt'" "video_subtitled.mp4"

---

Merge/Concatenate Videos

Join multiple videos together.

User might say: "merge video1 and video2", "combine clips", "join intro and main"

Method: First create a text file listing videos, then concatenate.

Step 1 - Create file list (files.txt):

file 'video1.mp4'
file 'video2.mp4'
file 'video3.mp4'

Step 2 - Concatenate:

ffmpeg -y -hide_banner -f concat -safe 0 -i files.txt -c copy "merged.mp4"

---

Time Format Reference

Use these formats for timestamps:

HH:MM:SS → 01:30:45 (1 hour 30 min 45 sec)
MM:SS → 05:30 (5 min 30 sec)
SS → 90 (90 seconds)
HH:MM:SS.mmm → 00:01:23.500 (with milliseconds)

---

Response Format

When generating commands:

Show the FFmpeg command in a code block
Briefly explain what it does
Mention if output filename was assumed

Example response:

Here's the command to cut your video from 1:21 to 1:35:

bash ffmpeg -y -hide_banner -i "video.mp4" -ss 00:01:21 -to 00:01:35 -c copy "video_trimmed.mp4"


This extracts the segment without re-encoding (using -c copy for speed). Output saved as video_trimmed.mp4.

Code Example

clawhub ffmpeg-video-editor "剪裁视频从 1 分钟到 5 分钟，转换为 1080p MP4，压缩质量为 75%"

Also popular in Video Editing

View all

Douyin Video Fetch

clawhubVideo Editing Easy

3.4

Download Douyin (TikTok China) videos to local files with watermark-free preference. Accepts URL or video_id, supports batch lists and a unified output directory — provides raw material for downstream video analysis or remixing.

clawhub install douyin-video-fetch

Video Clip

clawhubVideo Editing Easy

3.3

Use when the user wants to trim, cut, or extract a specific segment from a video by time range — e.g. "cut from 1:30 to 3:00", "trim the first 2 minutes", "e...

clawhub install video-clip

Video Translator

clawhubVideo Editing Medium

3.2

Real time video translation / dubbing skill. Translate user-provided video (file or URL) and return preview_url. 适用于视频直译、视频翻译、视频配音、字幕翻译出片。

clawhub install video-translator

Video Translate

clawhubVideo Editing Medium

3.2

Translate and dub existing videos into multiple languages using HeyGen. Use when: (1) Translating a video into another language, (2) Dubbing video content wi...

clawhub install video-translate

Remotion Video Toolkit

clawhubVideo Editing Easy Tested

9.5

Complete toolkit for programmatic video creation with Remotion + React.

7017K

clawhub install shreefentsar/remotion-video-toolkit

FFmpeg

clawhubVideo Editing Medium Tested

3.6

Process video and audio with correct codec selection, filtering, and encoding settings.

1.5K

clawhub install ffmpeg

Test Report

Real cold-start runs of this skill on Claude Code and Codex.

3 tasksClaude Code · Codex100% pass

fve1

Compress a video from a plain-English request

Input: Plain-English brief on an 18.7 MB 1080p clip: "This video is too big to share — compress it so it’s small enough to send easily (aim for a few MB), but keep it looking good and don’t drop below 720p." Save as compressed.mp4.

Skill: FFmpeg Video EditorAlso used: ffmpeg

Claude CodePass

Score /5

Interpreted "few MB + keep quality + ≥720p" as a CRF 23 (visually transparent) H.264 re-encode, downscaled 1080p→720p, 128k AAC + faststart — one clean encode: `-vf scale=-2:720 -c:v libx264 -crf 23 -preset slow -c:a aac -b:a 128k -movflags +faststart`.

CodexPass

Score /5

Rounds

64s

Time

Probed input (ffprobe), then the SAME recipe (CRF 23 / 720p / faststart / 128k AAC) with explicit `-map 0:v:0 -map 0:a?` stream mapping; one retry to quote the `0:a?` glob, then verified size/resolution/duration with ffprobe + stat. Byte-identical output. 26.2k tokens.

codex-cli 0.140

Output: compressed.mp4 — 1280×720, 10.0s, 2.02 MB (from 18.7 MB → 9.3× smaller). Both agents byte-identical (md5 271dac81…).

Verdict: Both correctly translated the plain-English brief into the professional recipe (CRF 23 visually-transparent, 720p floor, faststart, 128k AAC) and converged on byte-identical 2.02 MB output (9.3×) — a strong endorsement of the skill. Deterministic-task pattern holds (equivalent output); the usual efficiency gap remains: Claude 4 actions / 46s vs Codex 10 actions / 64s (incl. one glob-quoting retry).

fve2

Reframe landscape → vertical 9:16 (TikTok/Shorts)

Input: Plain-English brief on a 1280×720 (16:9) clip: "Make this video vertical for TikTok / Shorts — 9:16, 1080×1920 — and keep it looking good (don’t just stretch it)." A real-world ambiguous task: crop-to-fill vs blurred-pad are both valid — the agent chooses.

Skill: FFmpeg Video EditorAlso used: ffmpeg

Claude CodePass

Score /5

Rounds

296s

Time

Chose **scale-to-fill + center-crop** — `scale=1080:1920:force_original_aspect_ratio=increase,crop=1080:1920,setsar=1`. Full-bleed 9:16, no bars, no distortion (SAR 1:1); trades off the source’s left/right edges. Cleanest for an immersive Short. (296s wall-clock is latency noise — 6 actions.)

claude-opus-4-8

fve3

Add a logo watermark (bottom-right, semi-transparent)

Input: Plain-English brief on a 640×360 clip + a transparent "skills.video" logo.png: "Add the skills.video logo as a watermark in the bottom-right corner — make it subtle (~65–70% opacity) with a small margin, scaled sensibly for this video." Save as watermarked.mp4.

Skill: FFmpeg Video EditorAlso used: ffmpeg, ImageMagick (logo)

Claude CodePass

Score /5

Rounds

41s

Time

One filter_complex: scaled the logo to 200×50, set ~67% opacity via `format=rgba,colorchannelmixer=aa=0.67`, overlaid bottom-right `overlay=W-w-12:H-h-12`. Frame-verified the wordmark is visible.

claude-opus-4-8

Environment: macOS · ffmpeg 8.1.1 / ffprobe · symmetric cold start (codex exec vs a fresh Claude Code subagent). Faithful run — this skill is a natural-language→FFmpeg-command translator (command-pattern knowledge, no binary to install); both agents interpreted the plain-English brief and produced the command. Deterministic — outputs byte-identical. No network.

Score /5

Rounds

61s

Time

Chose **fit + blurred-background pad** — the whole landscape frame centered over a blurred scale-to-fill backdrop. Preserves ALL content (nothing cropped) with a stylized border; the actual video occupies less vertical space. Both 6 actions; Codex faster on wall-clock (61s).

Output: Both: valid 1080×1920 / 9:16 / 6s vertical MP4, no distortion — but DIFFERENT strategies (Claude crop-to-fill 1.97 MB; Codex blur-pad 2.74 MB). Shown here web-optimized to 720×1280.

Verdict: Both nailed the technical target (1080×1920, exact 9:16, no stretch) but made DIFFERENT, equally-valid design choices — the two canonical vertical-reframe strategies (crop-to-fill vs blur-pad). A clean creative-split data point on an ambiguous real-world task: neither is wrong, it’s a content-dependent judgment call. Action count tied (6 each); Codex faster on wall-clock here.

Score /5

Rounds

67s

Time

Same technique — `colorchannelmixer=aa=0.68`, overlay bottom-right with a 16px margin, but scaled the logo smaller (120px wide, more restrained). Re-encoded at CRF 18, frame-verified. 7 shell steps.

Output: watermarked.mp4 — 640×360 / 6s with a semi-transparent "skills.video" wordmark in the bottom-right (Claude logo ~200px; Codex ~120px). Both correct; sizing/margin differ slightly. Shown web-optimized.

Verdict: Both correctly translated the plain-English brief into the right overlay recipe (scale → alpha via colorchannelmixer → corner overlay with margin) — near-convergence, differing only in logo size/margin (a taste call). A clean "涉及的其他 skills" touch: the watermark logo itself was made with ImageMagick. Efficiency close (6 vs 7 actions).

ffmpeg-video-editor

Install

🤖Use in AI Agents

Claude Code

OpenAI Codex / Cursor / Windsurf

OpenClaw Ecosystem

Environment & Dependencies

SKILL.md

How to Generate Commands

Command Reference

Cut/Trim Video

Format Conversion

Change Aspect Ratio

Change Resolution

Compress Video

Extract Audio

Remove Audio

Change Speed

Convert to GIF

Rotate/Flip Video

Extract Screenshot/Frame

Add Watermark/Logo

Burn Subtitles

Merge/Concatenate Videos

Time Format Reference

Response Format

Code Example

Also popular in Video Editing

Douyin Video Fetch

Video Clip

Video Translator

Video Translate

Remotion Video Toolkit

FFmpeg

Test Report

Compress a video from a plain-English request

Reframe landscape → vertical 9:16 (TikTok/Shorts)

Add a logo watermark (bottom-right, semi-transparent)

Use in AI Agents