A high-performance AI agent skill designed to summarize, translate, and extract deep insights from YouTube videos using a transcript-first approach.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install youtube-digest
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install youtube-digest using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is YouTube Digest?
YouTube Digest is a sophisticated tool within the ecosystem of Openclaw Skills that enables AI agents to process video content with surgical precision. By prioritizing manual subtitles and structured metadata, the skill allows developers and researchers to bypass manual watching and jump straight to the core insights of any video.
Whether you need a Chinese translation of a technical talk or a bulleted timeline of a long-form interview, this skill provides the necessary automation. It minimizes bandwidth by fetching transcripts first and only downloading full video assets when advanced visual grounding or frame extraction is required via ffmpeg.
YouTube Digest Use Cases
- Rapidly generating multi-language summaries for long-form educational videos.
- Extracting accurate transcripts and notable quotes for research and documentation.
- Creating automated chapter notes and timestamped timelines for content creators.
- Performing visual analysis and key frame extraction for video content verification.
- Building searchable knowledge bases from video-heavy data sources.
How YouTube Digest Works
- The agent receives a YouTube URL and executes the fetch script to collect metadata and available subtitles.
- It prioritizes manual subtitles over auto-generated ones to ensure the highest possible text accuracy.
- The skill parses the summary.json and transcript.txt files to build a contextual understanding of the video.
- If visual analysis is requested, it utilizes ffmpeg to extract specific key frames for grounding the AI's response.
- Finally, it outputs a structured response including an executive summary, key insights, and actionable timestamps.
YouTube Digest Setup
To integrate this component of Openclaw Skills, ensure you have the required runtimes installed, then use the provided fetch script.
# Ensure dependencies are available
# Requires yt-dlp, deno (for 2026+ yt-dlp), and ffmpeg
# Basic video extraction
python3 scripts/fetch_youtube.py "<youtube-url>" --out /tmp/youtube-digest
# Extraction with proxy and specific language preference
python3 scripts/fetch_youtube.py "<youtube-url>" --langs zh.*,en.* --proxy http://your-proxy:port --out /tmp/youtube-digest
YouTube Digest Data Schema & Taxonomy
The skill organizes its output into a dedicated directory for clean data management:
| File | Description |
|---|---|
summary.json |
Contains title, uploader, duration, and subtitle extraction status. |
transcript.txt |
The flattened, plain-text version of the video transcript. |
*.vtt / *.srt |
The raw subtitle files fetched directly from YouTube. |
frames/ |
(Optional) Directory containing extracted key frames for visual grounding. |
YouTube Digest Advanced Features
- Advanced proxy support to bypass 429 rate limits and bot detection.
- Flexible language filtering that prefers Chinese and English subtitles for global accessibility.
- Visual grounding capabilities through ffmpeg-driven key frame extraction.
- Intelligent failure handling that distinguishes between manual and noisy auto-generated transcripts.
- Multi-agent compatibility for transcript-first research workflows within Openclaw Skills.
SKILL.md
Loading
Related Openclaw Skills
Family Rebate Manager
A unified family shopping reward management tool that aggregates rebate accounts and tracks savings for the entire household.
Xiaoxiyouxi 5min Report
An automated intelligence tool that scrapes Tencent data to generate concise, revenue-focused WeChat mini game industry reports delivered directly to Feishu.
AI Script Kill (AI 剧本杀)
A sophisticated turn-based Isekai life simulation engine that allows players to experience an entire lifetime in a procedurally generated world.
Data Research Starter
A structured research initiation skill that converts ambiguous analysis requests into clear, verified research outlines through guided dialogue.
Flight Requirement Review Agent
An expert AI agent designed to conduct structured audits and scoring for frontend, backend, and operational flight product requirement documents.
Shopping Price Comparison (Safe Mode)
A privacy-first shopping assistant that aggregates product pricing from major e-commerce platforms using public search data to generate detailed comparison reports.
Featured*
