AI-Powered YouTube Clipper Trimming Bilingual Subtitles

AI‑Powered YouTube Clipper: Trim, Translate, & Share in Minutes

If you’ve ever needed a quick, high‑quality clip from a long YouTube video—and wanted bilingual subtitles baked right in—this open‑source skill is exactly what you need.

Why This Project Stands Out

Feature What It Means Why It Matters
Semantic Chaptering Uses Claude to read the subtitle script and carve the video into meaningful 2‑5 minute chapters, not just straight‑time splits. You get context‑based chapters that match real topic shifts, making it easier to choose the best clip.
Frame‑Accurate Clipping Employs FFmpeg to cut video at precise frames. No half‑second slippage means cleaner starts/ends.
Batch Translation Translates up to ~25 subtitles at once, slashing API calls by 95%. Faster, cheaper, and more consistent translations.
Bilingual & Buried Subtitles Generates a two‑language SRT, and can hard‑code the subtitles with custom styling. Perfect for social‑media‑ready clips that look polished.
Summary Generation Produces short, platform‑friendly texts for Sina Weibo, TikTok, WeChat, etc. Ready‑to‑post captions save you time.
One‑Line Install npx skills add https://github.com/op7418/Youtube-clipper-skill Removes the need to manually clone or set up.

Getting Started

1️⃣ Install the Skill

# Recommended: npx
npx skills add https://github.com/op7418/Youtube-clipper-skill

The script copies the repo into ~/.claude/skills/youtube-clipper/, installs Python deps (yt-dlp, pysrt, python-dotenv) and checks for ffmpeg with libass support.

Mac users: if you have the default ffmpeg from Homebrew, uninstall it (brew uninstall ffmpeg) and install the full build (brew install ffmpeg-full).

2️⃣ Configure Environment Variables

Edit ~/.claude/skills/youtube-clipper/.env:

# Where FFmpeg lives – empty means auto‑detect
FFMPEG_PATH=

# Output directory – defaults to current folder
OUTPUT_DIR=./youtube-clips

# Max video quality
MAX_VIDEO_HEIGHT=1080

# Batch size for translation
TRANSLATION_BATCH_SIZE=20

# Translation target language
TARGET_LANGUAGE=中文

# Desired chapter length (seconds – 180 to 300 recommended)
TARGET_CHAPTER_DURATION=180

Use a .env file for simple overrides—no hard‑coding required.

3️⃣ Run a Demo

# Ask Claude to clip a video
Clip this YouTube video: https://youtube.com/watch?v=VIDEO_ID

Claude will: 1. Download the video & English subtitles. 2. Analyze the subtitle script and suggest semantic chapters. 3. Let you pick which chapters to keep. 4. Clip, translate, burn subtitles, and summarize.

The outputs land under ./youtube-clips/<timestamp>/: * Chapter_Title_clip.mp4 – raw clip * Chapter_Title_with_subtitles.mp4 – hardcoded subtitles * Chapter_Title_bilingual.srt – two‑language subtitle file * Chapter_Title_summary.md – social‑media copy

Advanced Use Cases

Scenario How to Execute Key Output
Highlight a tech interview Same as demo – pick the desired chapters (e.g., 00:00‑03:15, 03:15‑06:30) Clean clips with subtitles for each segment
Create course snippets Add flags: Generate bilingual subtitles: Yes, Burn subtitles into video: Yes, Generate summary: Yes Ready‑to‑publish short lessons
Fast, offline processing Set YT_DLP_PROXY if behind a corporate proxy; adjust FFMPEG_PATH if locally installed Reliable runs on restricted networks

Troubleshooting Quick‑Fixes

Issue Symptom Fix
FFmpeg subtitle burning fails Option not found: subtitles Install ffmpeg-full or libass-dev (sudo apt install ffmpeg libass-dev)
Video download slow Low transfer speeds Set YT_DLP_PROXY in .env
Subtitle translation errors Rate limiting or network hiccups Reduce TRANSLATION_BATCH_SIZE, check network connection
Broken output filenames Special characters in chapter titles Skill sanitizes titles, but ensure titles are concise

Why Adopt It?

  • Zero Cost – MIT‑licensed, no hidden fees.
  • Community‑Ready – Fork to tweak chapter algorithm, translation provider, etc.
  • Rapid Prototyping – Developers can prototype AI‑driven media workflows.
  • Social‑Media‑Optimized – Ready‑made captions & bilingual subtitles.

Get Involved

Feel free to contribute via Pull Requests: improve regex, add new subtitle styles, or integrate other translation APIs. Open issues for bugs or feature ideas.


The YouTube Clipper Skill turns a 1‑hour YouTube lecture into a set of polished clips in minutes—ready for Instagram Reels, TikTok short‑clips, or internal training videos. Grab it today and supercharge your video editing workflow.

Original Article: View Original

Share this article