菜单
Analyze raw video content using Gemini to identify speakers, topics, key moments, and potential clip opportunities
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Generate karaoke-style word-level timestamps by aligning script text to audio using Qwen3-ForcedAligner + jieba for Chinese word segmentation. Use when the user says 'align captions', 'karaoke timestamps', 'word timestamps', 'caption alignment', 'sync text to audio'.
Audio processing utilities - noise reduction, normalization, enhancement
Extract a video segment using FFmpeg with precise start/end times
Find naturally clean, coherent video segments worth keeping (selection over repair)
Text-guided audio source separation using SAM-Audio via mlx-audio
ASR with ~30ms timestamp precision using Qwen3-ASR + ForcedAligner
| name | analyze-video |
| description | Analyze raw video content using Gemini to identify speakers, topics, key moments, and potential clip opportunities |
Use this skill when you need to understand the content of a video file before creating clips.
Sends a video to Gemini 3 Pro for comprehensive analysis, extracting:
# Analyze a single video
python skills/analyze-video/analyze.py <video_path>
# Output goes to assets/outputs/analysis/<video_name>.json
{
"video_file": "path/to/video.mp4",
"duration_seconds": 120,
"speakers": [
{"name": "Speaker Name", "role": "Title/Organization", "first_seen": "00:05"}
],
"topics": ["Topic 1", "Topic 2"],
"notable_quotes": [
{"speaker": "Name", "quote": "...", "timestamp": "01:23", "context": "..."}
],
"visual_highlights": [
{"timestamp": "02:15", "description": "...", "type": "gesture|reveal|reaction"}
],
"clip_opportunities": [
{
"start": "01:20",
"end": "01:45",
"type": "quote|exchange|highlight",
"description": "...",
"score": 8
}
]
}