watch

Transcribe YouTube videos via Gemini and index results into Oracle.

116|56|Updated Jan 18, 2026
One-click install
npx skills add https://github.com/Soul-Brews-Studio/oracle-skills-cli --skill watch-soul-brews-studio
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: watch
Source: https://github.com/Soul-Brews-Studio/oracle-skills-cli/tree/main/src/skills/watch
Command: npx skills add https://github.com/Soul-Brews-Studio/oracle-skills-cli --skill watch-soul-brews-studio

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires yt-dlp, mosquitto, and includes scripts (resource) components.

What problem does it solve?

This skill automates YouTube video transcription by sending content to Gemini for transcription and then indexing results into Oracle, enabling quick knowledge capture from video content.

Core Features & Use Cases

  • Automated Transcription: Transcribe YouTube videos with timestamps using Gemini.
  • Knowledge Extraction & Indexing: Save transcripts as learning files under ψ/memory/learnings and index to Oracle for fast search.
  • Use Case: A researcher can watch a long lecture, have Gemini generate a structured document, and store key takeaways for later queries.

Quick Start

Use the watch skill to transcribe a YouTube video:

  • Transcribe: bun src/skills/watch/scripts/transcribe.ts https://youtube.com/watch?v=xxx
  • Get metadata: bun src/skills/watch/scripts/get-metadata.ts https://youtube.com/watch?v=xxx
  • Get captions: bun src/skills/watch/scripts/get-cc.ts https://youtube.com/watch?v=xxx en
  • Save learning: bun src/skills/watch/scripts/save-learning.ts "Video Title" https://youtube.com/watch?v=xxx <VIDEO_ID> "<transcript>"

Frequently Asked Questions about watch

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

FAQPage Schema
How do I transcribe a YouTube video for research and knowledge indexing?

You can extract YouTube captions by running the get-cc script with the video URL and language code. This retrieves existing captions before sending the content to Gemini for full transcription and analysis.

Do I need yt-dlp and MQTT to automate YouTube video transcription?

Yes, the automated transcription workflow relies on yt-dlp to fetch video content and MQTT integration to communicate with Gemini. You also need bun installed to execute the TypeScript scripts.

What is the best way to generate structured documents from long video lectures?

The best way to create structured documents from video lectures is using the watch Skill's research mode, which utilizes Gemini to transcribe and analyze the content, storing key takeaways as indexed learning files.

Can I retrieve YouTube video metadata and transcripts separately before processing?

Yes, you can retrieve video metadata using the get-metadata script and fetch captions independently with the get-cc script. This allows you to verify video details before committing to full Gemini transcription and indexing.

What formats and outputs are supported when saving transcribed video learnings?

Transcribed video learnings are saved as text files under the memory/learnings directory with optional captions. The Skill supports chat, research, and canvas output modes for different types of video analysis workflows.