video-to-note

Name: video-to-note
Author: LjyYano

38stars3forks

How to install video-to-note?

ljyyano/skill-pack/video-to-note

$npx -y skills add LjyYano/skill-pack --skill video-to-note

Installs into the current project.

Summary

Use when the user provides a YouTube or Bilibili URL and wants it summarized into an Obsidian note. Prefers auto-generated subtitles over ASR. Uses Obsidian CLI for note creation.

Files · 1

View on GitHub

SKILL.md

1	# Video → Obsidian Note
2
3	## Overview
4
5	Get transcript from video subtitles (preferred) or Qwen ASR fallback (Alibaba Cloud), then write a structured Obsidian note via the `obsidian` CLI.
6
7	Supported platforms: YouTube, Bilibili
8
9	Output directory:
10	- YouTube → `AI/YouTube/`
11	- Bilibili → `AI/Bilibili/`
12
13	---
14
15	## Prerequisites
16
17	- `yt-dlp` — subtitle download & audio download (supports YouTube & Bilibili)
18	- `ffmpeg` — audio conversion & splitting (ASR fallback only)
19	- `python3` + `requests` — API calls (ASR fallback only)
20	- `ALIYUN_API_KEY` — DashScope API Key for Qwen ASR (ASR fallback only)
21	- `obsidian` CLI — note creation (Obsidian must be running)
22	- Bilibili cookies (optional) — for member-only or age-restricted Bilibili videos, use `--cookies-from-browser chrome`
23
24	---
25
26	## Step-by-Step Workflow
27
28	### 0. Detect platform
29
30	```python
31	import re
32	url = "URL"
33	if re.match(r'https?://(www\.)?(bilibili\.com\|b23\.tv)', url):
34	platform = "bilibili"
35	output_dir = "AI/Bilibili"
36	elif re.match(r'https?://(www\.)?(youtube\.com\|youtu\.be)', url):
37	platform = "youtube"
38	output_dir = "AI/YouTube"
39	else:
40	print("Unsupported platform")
41	exit(1)
42	```
43
44	> Use `platform` and `output_dir` throughout the workflow.
45
46	### 1. Get video metadata + detect subtitles
47
48	```bash
49	yt-dlp --dump-json --skip-download "URL" 2>&1 \| python3 -c "
50	import json, sys
51	for line in sys.stdin.read().strip().split('\n'):
52	try:
53	d = json.loads(line)
54	print('TITLE:', d.get('title',''))
55	# Bilibili uses 'uploader', YouTube uses 'channel'
56	print('CHANNEL:', d.get('channel','') or d.get('uploader',''))
57	print('UPLOAD_DATE:', d.get('upload_date',''))
58	print('DURATION:', d.get('duration_string',''))
59	print('VIEW_COUNT:', d.get('view_count',''))
60	print('DESCRIPTION:', d.get('description','')[:500])
61	# Detect auto-generated subtitles
62	auto_subs = d.get('automatic_captions', {})
63	manual_subs = d.get('subtitles', {})
64	if auto_subs:
65	langs = list(auto_subs.keys())
66	print('AUTO_SUBS_AVAILABLE:', ','.join(langs))
67	if manual_subs:
68	langs = list(manual_subs.keys())
69	print('MANUAL_SUBS_AVAILABLE:', ','.join(langs))
70	if not auto_subs and not manual_subs:
71	print('NO_SUBS_AVAILABLE')
72	break
73	except: continue
74	"
75	```
76
77	> Bilibili note: If download fails (member-only / age-restricted), retry with `--cookies-from-browser chrome`.
78
79	Subtitle decision logic:
80
81	\| Condition \| Action \|
82	\|-----------\|--------\|
83	\| `AUTO_SUBS_AVAILABLE` contains `en` or `zh-Hans` or `zh` \| Download subtitle → go to Step 2A \|
84	\| `MANUAL_SUBS_AVAILABLE` contains `en` or `zh-Hans` or `zh` \| Download subtitle → go to Step 2A \|
85	\| Any auto/manual subs available \| Download the first available language → go to Step 2A \|
86	\| `NO_SUBS_AVAILABLE` \| Fall back to ASR → go to Step 2B \|
87
88	> Priority: Prefer `zh-Hans` > `zh` > `en` > first available language.
89
90	### 2A. Download subtitle (preferred path)
91
92	```bash
93	# Use best subtitle format (srv3/vtt/srt), yt-dlp auto-selects best
94	yt-dlp --write-auto-sub --sub-lang LANG --sub-format srv3/vtt/srt \
95	--skip-download -o "./.yt_sub_%(id)s" "URL" 2>&1 \| tail -3
96	```
97
98	Then parse the subtitle file to extract clean text with logical paragraph segmentation (based on timestamp gaps > 3s):
99
100	```python
101	import re
102
103	def _parse_timestamp(ts_str):
104	"""Parse SRT/VTT timestamp to seconds."""
105	# Handle both SRT (00:01:23,456) and VTT (00:01:23.456) formats
106	ts_str = ts_str.strip().replace(',', '.')
107	parts = ts_str.split(':')
108	if len(parts) == 3:
109	h, m, s = parts
110	return int(h) * 3600 + int(m) * 60 + float(s)
111	elif len(parts) == 2:
112	m, s = parts
113	return int(m) * 60 + float(s)
114	return 0.0
115
116	def _segment_by_gaps(cues, gap_threshold=3.0):
117	"""Group cues into paragraphs based on timestamp gaps.
118
119	When the gap between consecutive cues exceeds `gap_threshold` seconds,
120	a new paragraph is started. This creates natural logical breaks.
121	"""
122	if not cues:
123	return ""
124	paragraphs = []
125	current_lines = [cues[0][1]]
126	for i in range(1, len(cues)):
127	gap = cues[i][0] - cues[i-1][0]
128	if gap > gap_threshold:
129	paragraphs.append(" ".join(current_lines))
130	current_lines = [cues[i][1]]
131	else:
132	current_lines.append(cues[i][1])
133	if current_lines:
134	paragraphs.append(" ".join(current_lines))
135	return "\n\n".join(paragraphs)
136
137	def parse_srt(srt_path):
138	"""Parse SRT file and return paragraph-segmented text."""
139	with open(srt_path, 'r', encoding='utf-8') as f:
140	content = f.read()
141	blocks = re.split(r'\n\n+', content.strip())
142	cues = [] # list of (end_time, text)

People who liked this also viewed

6 picks

Type

lark-drive飞书云空间（云盘/云存储）：管理 Drive 文件和文件夹，包含上传/下载、创建文件夹、复制/移动/删除、查看元数据、评论/权限/订阅、标题、版本、飞书文档密级标签（secure labels）和本地文件导入。用户需要整理云盘目录、处理云空间资源 URL/token、判断链接类型/真实 token/标题，或导入 Word/Markdown/Excel/CSV/PPTX/.base 为 docx/sSkillsJul 2026382k16k
lark-sharedUse for lark-cli setup/auth tasks: auth login/status/logout, user vs bot identity, business-domain permissions (--domain, including all/docs/drive), missing scopes, revoking authorization, or handlingSkillsJul 2026382k16k
lark-whiteboard飞书画板：查询和编辑飞书云文档中的画板。支持导出画板为预览图片、导出原始节点结构、使用多种格式更新画板内容。当用户需要查看画板内容、导出画板图片、编辑画板时使用此 skill。不负责：飞书云文档内容编辑（lark-doc）、文档内嵌电子表格/Base（lark-sheets / lark-base）。SkillsJul 2026380k16k
lark-task飞书任务：管理任务、清单和任务智能体。创建待办任务、查看和更新任务状态、拆分子任务、组织任务清单、分配协作成员、上传任务附件、注册或注销任务智能体、更新任务智能体的主页数据、写入智能体任务记录。当用户需要创建待办事项、查看任务列表、跟踪任务进度、管理项目清单或给他人分配任务、为任务上传附件文件、注册注销任务智能体、更新智能体主页数据、写入任务记录时使用。SkillsJul 2026380k16k
lark-attendance飞书考勤打卡：查询自己的考勤打卡记录SkillsJul 2026330k16k
lark-apps妙搭（Spark/Miaoda）应用开发与托管：应用创建、本地全栈开发、云端生成迭代、创意设计（UI mockup / 可交互原型 / 线框图 / 落地页 / 仪表盘 / 幻灯片 deck / 视觉探索）、AI相关能力和飞书平台能力或者其他外部能力集成、日志/Trace/监控指标/PV/UV 查询、环境变量管理、应用角色与成员管理、自动化触发器（定时/记录变更/Webhook/飞书审批）。当用户要SkillsJul 2026227k16k

video-to-note

38stars3forks

View on GitHub

How to install video-to-note?

ljyyano/skill-pack/video-to-note

$npx -y skills add LjyYano/skill-pack --skill video-to-note

Installs into the current project.

Summary

Use when the user provides a YouTube or Bilibili URL and wants it summarized into an Obsidian note. Prefers auto-generated subtitles over ASR. Uses Obsidian CLI for note creation.

Files · 1

View on GitHub

SKILL.md

1	# Video → Obsidian Note
2
3	## Overview
4
5	Get transcript from video subtitles (preferred) or Qwen ASR fallback (Alibaba Cloud), then write a structured Obsidian note via the `obsidian` CLI.
6
7	Supported platforms: YouTube, Bilibili
8
9	Output directory:
10	- YouTube → `AI/YouTube/`
11	- Bilibili → `AI/Bilibili/`
12
13	---
14
15	## Prerequisites
16
17	- `yt-dlp` — subtitle download & audio download (supports YouTube & Bilibili)
18	- `ffmpeg` — audio conversion & splitting (ASR fallback only)
19	- `python3` + `requests` — API calls (ASR fallback only)
20	- `ALIYUN_API_KEY` — DashScope API Key for Qwen ASR (ASR fallback only)
21	- `obsidian` CLI — note creation (Obsidian must be running)
22	- Bilibili cookies (optional) — for member-only or age-restricted Bilibili videos, use `--cookies-from-browser chrome`
23
24	---
25
26	## Step-by-Step Workflow
27
28	### 0. Detect platform
29
30	```python
31	import re
32	url = "URL"
33	if re.match(r'https?://(www\.)?(bilibili\.com\|b23\.tv)', url):
34	platform = "bilibili"
35	output_dir = "AI/Bilibili"
36	elif re.match(r'https?://(www\.)?(youtube\.com\|youtu\.be)', url):
37	platform = "youtube"
38	output_dir = "AI/YouTube"
39	else:
40	print("Unsupported platform")
41	exit(1)
42	```
43
44	> Use `platform` and `output_dir` throughout the workflow.
45
46	### 1. Get video metadata + detect subtitles
47
48	```bash
49	yt-dlp --dump-json --skip-download "URL" 2>&1 \| python3 -c "
50	import json, sys
51	for line in sys.stdin.read().strip().split('\n'):
52	try:
53	d = json.loads(line)
54	print('TITLE:', d.get('title',''))
55	# Bilibili uses 'uploader', YouTube uses 'channel'
56	print('CHANNEL:', d.get('channel','') or d.get('uploader',''))
57	print('UPLOAD_DATE:', d.get('upload_date',''))
58	print('DURATION:', d.get('duration_string',''))
59	print('VIEW_COUNT:', d.get('view_count',''))
60	print('DESCRIPTION:', d.get('description','')[:500])
61	# Detect auto-generated subtitles
62	auto_subs = d.get('automatic_captions', {})
63	manual_subs = d.get('subtitles', {})
64	if auto_subs:
65	langs = list(auto_subs.keys())
66	print('AUTO_SUBS_AVAILABLE:', ','.join(langs))
67	if manual_subs:
68	langs = list(manual_subs.keys())
69	print('MANUAL_SUBS_AVAILABLE:', ','.join(langs))
70	if not auto_subs and not manual_subs:
71	print('NO_SUBS_AVAILABLE')
72	break
73	except: continue
74	"
75	```
76
77	> Bilibili note: If download fails (member-only / age-restricted), retry with `--cookies-from-browser chrome`.
78
79	Subtitle decision logic:
80
81	\| Condition \| Action \|
82	\|-----------\|--------\|
83	\| `AUTO_SUBS_AVAILABLE` contains `en` or `zh-Hans` or `zh` \| Download subtitle → go to Step 2A \|
84	\| `MANUAL_SUBS_AVAILABLE` contains `en` or `zh-Hans` or `zh` \| Download subtitle → go to Step 2A \|
85	\| Any auto/manual subs available \| Download the first available language → go to Step 2A \|
86	\| `NO_SUBS_AVAILABLE` \| Fall back to ASR → go to Step 2B \|
87
88	> Priority: Prefer `zh-Hans` > `zh` > `en` > first available language.
89
90	### 2A. Download subtitle (preferred path)
91
92	```bash
93	# Use best subtitle format (srv3/vtt/srt), yt-dlp auto-selects best
94	yt-dlp --write-auto-sub --sub-lang LANG --sub-format srv3/vtt/srt \
95	--skip-download -o "./.yt_sub_%(id)s" "URL" 2>&1 \| tail -3
96	```
97
98	Then parse the subtitle file to extract clean text with logical paragraph segmentation (based on timestamp gaps > 3s):
99
100	```python
101	import re
102
103	def _parse_timestamp(ts_str):
104	"""Parse SRT/VTT timestamp to seconds."""
105	# Handle both SRT (00:01:23,456) and VTT (00:01:23.456) formats
106	ts_str = ts_str.strip().replace(',', '.')
107	parts = ts_str.split(':')
108	if len(parts) == 3:
109	h, m, s = parts
110	return int(h) * 3600 + int(m) * 60 + float(s)
111	elif len(parts) == 2:
112	m, s = parts
113	return int(m) * 60 + float(s)
114	return 0.0
115
116	def _segment_by_gaps(cues, gap_threshold=3.0):
117	"""Group cues into paragraphs based on timestamp gaps.
118
119	When the gap between consecutive cues exceeds `gap_threshold` seconds,
120	a new paragraph is started. This creates natural logical breaks.
121	"""
122	if not cues:
123	return ""
124	paragraphs = []
125	current_lines = [cues[0][1]]
126	for i in range(1, len(cues)):
127	gap = cues[i][0] - cues[i-1][0]
128	if gap > gap_threshold:
129	paragraphs.append(" ".join(current_lines))
130	current_lines = [cues[i][1]]
131	else:
132	current_lines.append(cues[i][1])
133	if current_lines:
134	paragraphs.append(" ".join(current_lines))
135	return "\n\n".join(paragraphs)
136
137	def parse_srt(srt_path):
138	"""Parse SRT file and return paragraph-segmented text."""
139	with open(srt_path, 'r', encoding='utf-8') as f:
140	content = f.read()
141	blocks = re.split(r'\n\n+', content.strip())
142	cues = [] # list of (end_time, text)

People who liked this also viewed

6 picks

Type