laniameda-youtube-digest

Purpose: Watch a YouTube video so Michael doesn't have to. Extract what's actionable — tools, prompts, workflows, "if X then Y" mappings — and save to KB. For AI creatorship tutorials: extract model-specific techniques, prompt structures, and optionally encode new knowledge into studio skills.

Engine: Supadata API (supadata skill). Full API reference: references/supadata.md. Key decision: use Transcript for narrated content; use Extract (vision) when prompts/settings are shown on screen but not spoken aloud (e.g. ComfyUI node graphs, Midjourney UI, on-screen prompt text).

Cross-reference: This is the skill that laniameda-x-post-digest calls when it encounters YouTube links in X posts.

Triggers

• YouTube URL pasted
• "digest this video"
• "watch this for me"
• "extract from this video"
• "what's useful in this"
• "learn from this"
• "what does this teach"