| name | noteking-pro |
| description | 全网最强视频/录音处理工具。将视频、录音、会议转化为精美图文 PDF 讲义、结构化会议纪要、思维导图、学习闪卡等多种格式。支持说话人分离、降噪增强、中英混合、30+ 平台视频链接。涵盖会议/课堂/访谈/灵感/娱乐全场景，23 种输出模板，本地优先+云端可选。
|
| version | 2.0.0 |
| metadata | {"openclaw":{"requires":{"env":"[Truncated]","bins":"[Truncated]"},"primaryEnv":"NOTEKING_LLM_API_KEY","emoji":"👑","homepage":"https://github.com/bcefghj/noteking","os":["macos","linux","windows"],"install":["[Truncated]"]}} |
NoteKing Pro - 全网最强视频/录音处理工具
Use this skill when the user wants to:
- Convert a video or audio recording into structured notes
- Generate meeting minutes from a recording
- Transcribe audio/video with speaker identification
- Create study materials from lectures
- Process interviews, brainstorm sessions, or podcasts
Core Capabilities
1. Online Video Processing (30+ platforms)
Convert any online video URL into structured notes.
2. Local Recording Processing (NEW)
Process local audio/video files with advanced features:
- Speaker diarization: Identify who said what in multi-speaker recordings
- Noise reduction: Clean up noisy recordings (3 levels)
- Multi-file merge: Combine multiple recording segments
- Language detection: Auto-detect language including Chinese-English mix
3. 23 Output Templates
| Template | Name | Best For |
|---|
brief | 简要总结 | Quick overview |
detailed | 详细学习笔记 | Systematic study |
mindmap | 思维导图 | Visual knowledge structure |
flashcard | 闪卡 | Spaced repetition |
quiz | 测验题 | Self-testing |
timeline | 时间线笔记 | Timestamped notes |
exam | 考试复习 | Exam preparation |
tutorial | 教程步骤 | Step-by-step extraction |
news | 新闻速览 | News summary |
podcast | 播客摘要 | Discussion summary |
xhs_note | 小红书笔记 | Social media note |
latex_pdf | LaTeX PDF | Professional lecture notes |
custom | 自定义 | User-defined prompt |
meeting_minutes | 会议纪要 | Meeting notes with action items |
lecture_notes | 课堂笔记 | Knowledge hierarchy |
interview | 访谈记录 | Q&A with insights |
brainstorm | 灵感记录 | Idea capture |
news_digest | 新闻摘要 | 5W1H structured digest |
exam_prep | 考试复习 | Flashcards + mock questions |
cornell_notes | 康奈尔笔记 | Cornell note method |
podcast_shownotes | 播客笔记 | Show notes with chapters |
entertainment |
Supported Platforms
- Bilibili (B站), YouTube, Douyin (抖音), Xiaohongshu (小红书)
- TikTok, Twitter/X, Instagram, Twitch, Vimeo
- 1800+ additional sites via yt-dlp
- Local files: MP4, MKV, AVI, MOV, MP3, WAV, FLAC, M4A, OGG, AAC
Dependencies
Required:
ffmpeg - Audio/video processingpython3 >= 3.11
Python packages (auto-installed):
openai - LLM API clientyt-dlp - Video downloadingrich - Terminal UI
Optional (for best experience):
funasr - Best Chinese ASR (WER 8.4%)faster-whisper - Local multilingual ASRpyannote.audio - Speaker diarizationnoisereduce - Noise reduction
Environment Variables
| Variable | Required | Description |
|---|
NOTEKING_LLM_API_KEY | Yes | API key for LLM (MiniMax/OpenAI/DeepSeek) |
NOTEKING_LLM_BASE_URL | No | Custom API base URL |
NOTEKING_LLM_MODEL | No | Model name (default: gpt-4o-mini) |
HF_TOKEN | No | HuggingFace token for pyannote speaker diarization |
NOTEKING_PROXY | No | Proxy URL for YouTube access from China |
BILIBILI_SESSDATA | No | Bilibili login cookie |
Usage
Process Online Video
cd /path/to/noteking
python -m cli run "<VIDEO_URL>" -t <TEMPLATE>
Process Local Recording (NEW)
python -m cli process meeting.mp4 -t meeting_minutes -c "产品周会" --speakers 4
python -m cli process lecture.mp3 -t lecture_notes --denoise 2
python -m cli process part1.mp4 part2.mp4 -t meeting_minutes
python -m cli process interview.wav -t interview --format markdown,srt,json
Transcribe Only
python -m cli transcribe recording.mp4
Denoise Audio
python -m cli denoise noisy.wav --level 2
Merge Files
python -m cli merge segment1.mp4 segment2.mp4 -o merged.mp4
Scene Types
| Scene | Template | Description |
|---|
| meeting | meeting_minutes | 会议纪要: 参会人/议题/决议/行动项 |
| lecture | lecture_notes | 课堂笔记: 知识点/公式/习题 |
| interview | interview | 访谈: Q&A/观点/立场 |
| brainstorm | brainstorm | 灵感: 想法/思维导图/行动 |
| news | news_digest | 新闻: 5W1H/引用/背景 |
| exam | exam_prep | 考试: 闪卡/模拟题/要点 |
| entertainment | entertainment | 娱乐: 高光/金句/推荐 |
Workflow
For Online Videos:
- Parse URL → Detect platform
- Extract subtitles → CC → ASR → Visual fallback
- Generate notes → Apply template via LLM
- Save output → Markdown + SRT
For Local Recordings:
- Preprocess → Validate, merge files, extract audio, denoise
- Transcribe → Auto-detect language, FunASR(Chinese)/Whisper(multilingual)
- Diarize → Identify speakers with pyannote-audio
- Generate → LLM with scene context and speaker labels
- Output → Multi-format (Markdown/PDF/SRT/VTT/JSON)
Notes for AI Agents
- For meetings: use
process command with meeting_minutes template and specify --speakers
- For lectures: use
lecture_notes template, optionally with --denoise
- For noisy recordings: increase
--denoise level (1=light, 2=medium, 3=heavy)
- For long recordings split into parts: pass all files to
process command
- Default to
meeting_minutes for recordings unless user specifies otherwise
- For quick overview, use
smart_summary template
- Check
ffmpeg and python3 are installed before processing
