media-hub
Compare original and translation side by side
🇺🇸
Original
English🇨🇳
Translation
ChineseMedia Hub
Media Hub
Unified audio and video processing, transcription, and understanding.
统一的音视频处理、转录与内容理解工具。
When to Use
适用场景
Use this skill when:
- Processing audio or video files for transcription or conversion
- Generating subtitles or speech-to-text output
- Extracting frames or clips from video content
- Performing batch media operations or format conversions
Do NOT use this skill if:
- The task is creating presentation slides → use pitch-deck-creator
- The task requires map or location data → use amap-navigator
- You only need simple file copying without media analysis
Typical triggers:
- 「视频转文字」「音频转录」「生成字幕」
- "transcribe video", "convert media format", "audio to text"
- "extract video frames", "batch process audio", "媒体转换"
当你需要以下操作时使用本技能:
- 处理音频或视频文件以进行转录或格式转换
- 生成字幕或speech-to-text输出
- 从视频内容中提取帧或片段
- 执行批量媒体操作或格式转换
请勿在以下场景使用本技能:
- 需要创建演示幻灯片 → 使用pitch-deck-creator
- 需要地图或位置数据 → 使用amap-navigator
- 仅需简单文件复制而无需媒体分析
典型触发词:
- 「视频转文字」「音频转录」「生成字幕」
- "transcribe video", "convert media format", "audio to text"
- "extract video frames", "batch process audio", "媒体转换"
Workflow
工作流程
Step 1: Ingest Media
步骤1:导入媒体
Accept input via:
- File path to local audio/video
- URL to downloadable media
- Existing media directory for batch processing
支持以下输入方式:
- 本地音视频文件路径
- 可下载媒体的URL
- 用于批量处理的现有媒体目录
Step 2: Select Operation
步骤2:选择操作
Choose the appropriate processing mode:
- Transcription — speech-to-text with speaker diarization (optional)
- Format Conversion — re-encode to mp3, mp4, wav, webm, etc.
- Frame Extraction — capture keyframes at intervals or timestamps
- Subtitle Generation — produce SRT/VTT from transcribed audio
- Content Understanding — generate summaries or topic extraction from media
选择合适的处理模式:
- 转录(Transcription) — 带说话人分离(可选)的speech-to-text功能
- 格式转换(Format Conversion) — 重新编码为mp3、mp4、wav、webm等格式
- 帧提取(Frame Extraction) — 按时间间隔或时间戳捕获关键帧
- 字幕生成(Subtitle Generation) — 根据音频转录生成SRT/VTT格式字幕
- 内容理解(Content Understanding) — 从媒体内容生成摘要或提取主题
Step 3: Configure Parameters
步骤3:配置参数
- Source language for transcription (auto-detect if not specified)
- Output format and quality settings
- Target directory for processed files
- Batch size (for bulk operations)
- 转录的源语言(未指定则自动检测)
- 输出格式和质量设置
- 处理后文件的目标目录
- 批量处理的任务数量(批量操作时)
Step 4: Execute Processing
步骤4:执行处理
Run the relevant pipeline and report progress:
Processing: filename.ext
- Transcoding: 45%
- Transcribing: 20%运行相关处理流程并报告进度:
Processing: filename.ext
- Transcoding: 45%
- Transcribing: 20%Step 5: Deliver Results
步骤5:交付结果
Output files to a structured directory:
media-output/
├── transcript.txt / .srt / .vtt
├── converted/
├── frames/
└── summary.md将输出文件保存到结构化目录:
media-output/
├── transcript.txt / .srt / .vtt
├── converted/
├── frames/
└── summary.mdGuardrails
约束规则
Anti-patterns
反模式
- NEVER process copyrighted content without user confirmation
- NEVER overwrite original source files; always output to a separate directory
- NEVER return extremely long raw transcripts without a summary or table of contents
- 未经用户确认,绝不处理受版权保护的内容
- 绝不覆盖原始源文件;始终输出到单独目录
- 绝不返回极长的原始转录文本而不附带摘要或目录
Output Constraints
输出限制
- Original files remain untouched
- Transcripts include timestamps when subtitles are requested
- Frame extractions include metadata (timestamp, resolution)
- Summaries are concise and structured
- 原始文件保持不变
- 当请求字幕时,转录文本需包含时间戳
- 提取的帧需包含元数据(时间戳、分辨率)
- 摘要需简洁且结构化
Error Handling
错误处理
- Unsupported format: inform user and suggest conversion options
- Large files: split into chunks or request shorter segments
- Low-quality audio: flag accuracy warnings in transcription output
- 不支持的格式:告知用户并建议转换选项
- 大文件:分割为块或请求较短片段
- 低质量音频:在转录输出中标记准确性警告
Related Skills
相关技能
-
- amap-navigator — Geotag or route media content when location context is needed
- amap-navigator — 当需要位置上下文时,为媒体内容添加地理标签或路线信息
About UniqueClub
关于UniqueClub
This skill is part of the UniqueClub toolkit.
🌐 https://uniqueclub.ai | 📂 https://github.com/wulaosiji/skills
本技能是UniqueClub工具集的一部分。
🌐 https://uniqueclub.ai | 📂 https://github.com/wulaosiji/skills