viral-clips

name	viral-clips
description	Run the full Cut-AI viral clips pipeline. Analyzes a VTT transcript with Gemini to find highlights, cuts video clips with English and Chinese subtitles, extracts per-clip transcripts, and drafts platform-specific content for 7 platforms.
argument-hint	[data-directory]
disable-model-invocation	true

You are running the Cut-AI viral clips pipeline. This is a 3-phase automation that takes a recording session directory and produces viral short-form clips with subtitles and platform-specific content drafts.

The user passed this data directory: $ARGUMENTS

Setup

1. Derive the session name from the directory basename (e.g., data/020725 -> session = 020725).
2. Auto-detect the .vtt transcript file and .mp4 video file inside the $ARGUMENTS directory using Glob. There should be exactly one of each. If there are multiple, ask the user which to use.
3. Set these variables for the rest of the pipeline:
   • VTT = path to the .vtt file
   • MP4 = path to the .mp4 file
   • SESSION = session name
   • HIGHLIGHTS = output/<SESSION>/highlights_<SESSION>.json
   • OUTPUT_DIR = output/<SESSION>

Phase 1: Analyze Transcript

First create the output directory, then run the Gemini analysis to identify highlight clips:

mkdir -p output/<SESSION> && source venv/bin/activate && python script/cut_ai.py analyze --transcript <VTT> --output <HIGHLIGHTS>

After the command completes:

• Display a summary table of all highlights found (number, title, time range, category, viral score).
• STOP and ask the user to review <HIGHLIGHTS>. Tell them:
  • "Review and edit <HIGHLIGHTS> if needed (adjust timestamps, remove/reorder clips, edit titles). Reply go when ready to proceed to Phase 2."
• Do NOT proceed until the user confirms.

Phase 2: Cut Video with Dual-Language Subtitles

Once the user confirms, generate English and Chinese subtitle clips:

source venv/bin/activate && python script/cut_ai.py clip_dual --video <MP4> --highlights <HIGHLIGHTS> --subtitles <VTT> --output <OUTPUT_DIR>

This produces:

• <OUTPUT_DIR>/English/ - clips with burned-in English subtitles
• <OUTPUT_DIR>/Chinese/ - clips with burned-in bilingual English + Chinese subtitles

After completion, report how many clips were generated in each directory.

Phase 3: Extract Transcripts & Draft Content

Step 3a: Extract per-clip transcripts

source venv/bin/activate && python script/extract_transcripts.py <HIGHLIGHTS> <VTT> --output <OUTPUT_DIR>/transcripts

Step 3b: Read transcripts and draft platform content

Read each transcript file from <OUTPUT_DIR>/transcripts/.

Before drafting, read these two reference documents for guidance:

• docs/viral_content_guide.md - use this as the guide for what makes content viral, hook patterns, emotional triggers, content archetypes, and scoring criteria
• docs/platform_content_tips.md - use this as the reference for platform-specific formatting rules, character limits, algorithm signals, and tone guidelines for all 7 platforms

Then, for every clip, draft content for all 7 platforms and compile everything into a single file: <OUTPUT_DIR>/video_content_drafts.md.

Use this exact structure for the drafts file:

# Video Content Drafts - Session <SESSION>
Generated: <current date>

## Clip 1: <Title>
**Time:** <start> - <end>
**Category:** <category>

### X / Twitter
<draft>

### LinkedIn
<draft>

### YouTube Shorts
<draft>

### TikTok
<draft>

### WeChat Video (微信视频号)
<draft>

### Xiaohongshu (小红书)
<draft>

### Bilibili (B站)
<draft>

## Clip 2: <Title>
...

Platform drafting rules

Apply the detailed guidelines from docs/platform_content_tips.md for each platform. Key requirements:

• X / Twitter, LinkedIn, YouTube Shorts, TikTok: Write in English
• WeChat Video, Xiaohongshu, Bilibili: Write entirely in Chinese
• Bilibili: NEVER use emoji in titles or descriptions
• Xiaohongshu: Title must be 20 Chinese characters or fewer. NEVER include links in the description. Instead end with "留言获得视频全部内容" and a CTA: "觉得有用就点赞收藏分享！"
• For each platform, follow its specific character limits, hook formulas, hashtag rules, and tone guidelines as documented in the platform tips file
• Use the viral content archetypes and hook tier patterns from docs/viral_content_guide.md to maximize engagement

Platform CTAs (required at the end of every draft)

Every draft MUST end with a platform-appropriate call-to-action:

• X / Twitter: "Like, repost, and bookmark if this was useful!"
• LinkedIn: Invite comments with a question, then: "Like and share to help others see it."
• YouTube Shorts: "Like, subscribe, and share if this was helpful!"
• TikTok: "Like and share if this helped! Follow for more AI tips."
• WeChat Video (微信视频号): "觉得有用就点赞、点喜欢、分享到朋友圈！"
• Xiaohongshu (小红书): "留言获得视频全部内容。觉得有用就点赞收藏分享！"
• Bilibili (B站): "觉得有帮助就一键三连支持一下！" (no emoji)

Final Summary

After writing video_content_drafts.md, display:

1. Total clips processed
2. Files generated (list all output paths)
3. A reminder that the user should review the drafts and customize before posting