skills.rest
yang0
yang0COMMUNITY
Explore all skills by this author →
GitHub

youtube-search

Searches YouTube videos, extracts top comments, and downloads avatars to JSON and image files.

Updated May 10, 2026
Search Tags:#yt-dlp#comment extraction#batch automation#dataset building#youtube search#avatar download
One-click install
npx skills add https://github.com/yang0/youtube-search --skill youtube-search-yang0
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: youtube-search
Source: https://github.com/yang0/youtube-search/tree/main
Command: npx skills add https://github.com/yang0/youtube-search --skill youtube-search-yang0

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

This Skill eliminates the manual work of finding relevant YouTube videos, extracting high-signal comments, and downloading creator avatars so you can quickly compile datasets for research, analysis, or content workflows.

Core Features & Use Cases

  • Batch YouTube Video Search: Run multiple queries, de-duplicate results, and sort by view count to surface the most watched videos for your topic.
  • High-Signal Comment Extraction: Fetch top or newest comments for a specific video and produce a structured comments.json including author metadata and avatar URLs.
  • Avatar Download at Multiple Resolutions: Download a manifest-controlled set of avatar images from the extracted comments, skipping already-downloaded files for idempotent runs.
  • Use Case: For example, you can identify a niche topic, collect top comments from a shortlist of videos, and download the avatar images for moderation, study, or dataset building.

Quick Start

Tell the AI to help you do intent calibration first using 2-3 concrete example videos, then run a search with bun scripts/search_videos.js using your calibrated keyword list and an output directory of your choice.

Frequently Asked Questions about youtube-search

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I extract YouTube comments and download commenter avatars in bulk for a dataset?▼

You can extract YouTube comments and download commenter avatars by running a Bun script with yt-dlp that fetches top or newest comments into a structured JSON file, then uses a manifest to download avatar images at multiple resolutions into your specified output directory.

What is the best way to batch search YouTube videos and sort them by view count?▼

Batch YouTube video searching is handled by running multiple queries through a Bun script that de-duplicates results and sorts them by view count, outputting a structured JSON dataset of discovered videos for your chosen topics.

Do I need Bun and yt-dlp to run YouTube comment extraction and avatar downloads?▼

Yes, you need the Bun runtime to execute the provided scripts and yt-dlp to handle the YouTube comment extraction, along with optional cookies for authenticated access when fetching data from restricted videos.

How does the avatar download process handle reruns and existing files?▼

The avatar download process is idempotent, meaning it skips already-downloaded image files by checking against a manifest, ensuring reruns only fetch missing avatars and preventing duplicate downloads across multiple executions.

Can I collect YouTube creator assets and topic data across multiple queries simultaneously?▼

Yes, the workflow supports running multiple search queries and processing a shortlist of videos to discover topics, mine high-signal comments, and collect creator avatar assets across all results concurrently.

What format are the extracted YouTube comments and search results saved in?▼

Extracted YouTube comments and search results are saved as JSON files under your user-specified output directory, containing author metadata, avatar URLs, and structured video data for dataset building.

Related Skills

Explore More Data & Analytics Skills→
crazynomadcrazynomad
27

youtube-downloader

Download videos and playlists from YouTube and 1000+ sites via yt-dlp.

Community
Advanced
aws-samplesaws-samples
90

file-processing

Process and analyze CSV, JSON, and text data files with automated cleaning and transformation.

Official
Advanced
lgoodcodelgoodcode

image-pipeline

Automate image upload processing, moderation, and CDN delivery with Sharp and Gemini.

Community
Advanced
itda-skillsitda-skills

itda-kosis

Search and retrieve historical Korean national statistical data from KOSIS.

Official
Intermediate