/install bilibili-digest
Bilibili Digest Skill
Extract Bilibili (B站) video/column content and transform it into structured, note-ready Markdown with AI-generated summaries, timestamped key points, chapter segmentation, and cross-video integration.
How It Works
This skill operates in a pipeline:
- Parse – Extract BV/column ID from any Bilibili URL (including
b23.tvshort links) - Fetch – Retrieve video metadata (title, author, duration, views), CC subtitles, and danmaku via Bilibili public APIs
- Segment – Auto-detect video chapters from timestamps or subtitle gaps
- Summarize – Call LLM (DeepSeek) to generate structured summary, key points, resources, and action steps
- Export – Output as Markdown (default), Obsidian, JSON, or prepare for Notion/Feishu
Fallback Strategy for Missing Subtitles
If a video has no CC subtitles, the skill degrades gracefully:
- Generate a summary from title + description + danmaku highlights
- Prompt the user to paste their own notes as enrichment
API Rate Limiting
- Minimum 1-second interval between Bilibili API calls
- Exponential backoff retry (max 3 attempts) on 429 / timeout
- Local cache (24h TTL) at
~/.openclaw/data/bilibili-digest/cache/
Usage
clawhub run bilibili-digest --url \x3Cbilibili-url> [options]
Options
| Option | Type | Default | Description |
|---|---|---|---|
--url |
string | required | Single Bilibili URL |
--urls |
json-array | — | Multiple URLs for batch processing |
--summary-mode |
enum | detailed |
minimal, overview, detailed, mindmap |
--include-transcript |
bool | false | Include full transcript text |
--include-danmaku |
bool | false | Include danmaku sentiment analysis |
--export-format |
enum | markdown |
markdown, json, obsidian, notion, feishu |
--output-dir |
string | ./bilibili-notes/ |
Output directory |
--cross-video-merge |
bool | false | Merge insights across multiple videos |
--language |
enum | zh-CN |
zh-CN, en-US |
Sample Prompts
1. Extract a single video into structured notes (most common)
clawhub run bilibili-digest --url "https://www.bilibili.com/video/BV1xx411c7mD"
# → Structured Markdown with title, author, key points (with timestamps), chapters, resources
2. Batch process a tutorial series with cross-video merge
clawhub run bilibili-digest \
--urls '["https://www.bilibili.com/video/BV1xx01","https://www.bilibili.com/video/BV1xx02"]' \
--cross-video-merge
# → Combined knowledge tree across multiple videos
3. Quick minimal summary for social sharing
clawhub run bilibili-digest --url "https://www.bilibili.com/video/BV1xx411c7mD" \
--summary-mode minimal
# → One-liner + 3-5 core bullet points
4. Export in Obsidian format with full transcript
clawhub run bilibili-digest --url "https://www.bilibili.com/video/BV1xx411c7mD" \
--summary-mode detailed --export-format obsidian --include-transcript
# → Obsidian-compatible note with WikiLinks, tags, and YAML frontmatter
5. Mindmap summary for quick review
clawhub run bilibili-digest --url "https://www.bilibili.com/video/BV1xx411c7mD" \
--summary-mode mindmap
# → Hierarchical mindmap with main topics, subtopics, and key connections,
# ideal for sharing or importing into mind-mapping tools (XMind, etc.)
First-Success Path
Step 1: Install → clawhub install bilibili-digest
Step 2: Run → clawhub run bilibili-digest --url "https://www.bilibili.com/video/BV1xx411c7mD"
Step 3: Receive → Structured Markdown note (\x3C30 seconds)
Step 4: Copy to note-taking app → Value achieved
Core Scripts
The scripts/ directory contains Python modules:
| File | Purpose |
|---|---|
parser.py |
Extract BV/CV IDs and expand b23.tv short links |
api.py |
Bilibili API client with rate limiting and retry |
subtitle.py |
CC subtitle extraction and cleanup |
danmaku.py |
Danmaku density detection and sentiment grouping |
segmenter.py |
Chapter detection from subtitle gaps and transition words |
summarizer.py |
LLM-based structured summary generation |
exporter.py |
Markdown / Obsidian / JSON / Notion / Feishu export |
cross_merge.py |
Cross-video knowledge merging |
__init__.py |
Package init |
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install bilibili-digest - After installation, invoke the skill by name or use
/bilibili-digest - Provide required inputs per the skill's parameter spec and get structured output
What is Bilibili Digest?
Extract, structure, and summarize Bilibili video/column content into structured notes with timestamps, key points, and chapter indexes. It is an AI Agent Skill for Claude Code / OpenClaw, with 49 downloads so far.
How do I install Bilibili Digest?
Run "/install bilibili-digest" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Bilibili Digest free?
Yes, Bilibili Digest is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Bilibili Digest support?
Bilibili Digest is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Bilibili Digest?
It is built and maintained by haidong (@harrylabsj); the current version is v1.0.0.