skills.rest
neverprepared
neverpreparedOFFICIAL
Explore all skills by this author →
GitHub

youtube-harvester

Extract YouTube transcripts and metadata with timestamped chunking for vector retrieval.

Updated Aug 23, 2026
Search Tags:#rag#youtube#metadata#ingestion#transcripts#chunking#harvesting
One-click install
npx skills add https://github.com/neverprepared/ink-bunny --skill youtube-harvester-neverprepared
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: youtube-harvester
Source: https://github.com/neverprepared/ink-bunny/tree/main/reflex/plugins/reflex/skills/youtube-harvester
Command: npx skills add https://github.com/neverprepared/ink-bunny --skill youtube-harvester-neverprepared

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

YouTube transcripts and metadata are often scattered and hard to assemble for knowledge bases; this skill automates extraction of transcripts (manual and auto-generated) and enriches with metadata for efficient ingestion into retrieval systems.

Core Features & Use Cases

  • Transcript extraction from YouTube videos (manual and auto-generated)
  • Timestamp-aware chunking for scalable context windows
  • Playlist and channel harvesting for batch processing
  • Metadata enrichment (title, channel, upload_date, duration, view_count, tags, description)

Quick Start

Provide a YouTube URL and a collection name to begin harvesting transcripts and enriched metadata for downstream search.

Frequently Asked Questions about youtube-harvester

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

FAQPage Schema
How do I extract YouTube transcripts and metadata for a RAG knowledge base?▼

You can harvest YouTube transcripts and metadata by providing a video URL to an automated pipeline that extracts text and enriches it with details like title and channel for vector-based retrieval systems.

Can I batch harvest transcripts from an entire YouTube playlist or channel?▼

Yes, batch harvesting supports processing entire YouTube playlists and channels to extract transcripts and enriched metadata efficiently for downstream search ingestion.

How does timestamp-aware chunking work for YouTube transcripts?▼

Timestamp-aware chunking splits extracted YouTube transcripts into manageable context windows using their time markers, ensuring scalable and precise retrieval for vector-based ingestion.

Do I need yt-dlp and youtube-transcript-api to extract auto-generated subtitles?▼

You need youtube-transcript-api to extract both manual and auto-generated transcripts, while yt-dlp handles metadata harvesting, with optional whisper available for transcription fallback.

What metadata is enriched during YouTube video ingestion for search?▼

Metadata enrichment during ingestion includes the video title, channel name, upload date, duration, view count, tags, and description, providing comprehensive context for vector search.

What is the best way to handle videos without manual transcripts during ingestion?▼

For videos lacking manual transcripts, the pipeline extracts auto-generated subtitles using youtube-transcript-api, and can optionally use whisper as a transcription fallback to ensure complete ingestion.

Related Skills

Explore More Data & Analytics Skills→
LuminaVaultLuminaVault

youtube-content

Convert YouTube video transcripts into summaries, threads, and blog posts.

Official
Intermediate
patrykkopycinskipatrykkopycinski
6

rag-pipeline

Set up a RAG pipeline with Elasticsearch for vector retrieval.

Community
Intermediate
vecearvecear

embedding-pipeline-builder

Build document embedding and retrieval pipelines for scalable semantic search.

Community
Advanced
antonyfmunozantonyfmunoz

yt_dlp

Extract media metadata, transcripts, and audio from YouTube and other sites.

Community
Intermediate