Mfrostbutter

Transcript To Knowledge — AI skill for Claude Code

AI community

Turn any YouTube video into agent-ready research markdown: transcript -> structured LLM breakdown -> optional deep technical dive.

How to install Transcript To Knowledge

This entry records only its repository, not the path inside it, so there is no exact command to give. Open Mfrostbutter/transcript-to-knowledge and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Transcript To Knowledge does

Turn any YouTube video into agent-ready research markdown: transcript -> structured LLM breakdown -> optional deep technical dive.

Alternatives in AI

Claude Code Design Guide — From Early Internet Design Patterns to AI Agent Implementation — A Deep Dive into Claude Code for Developers 762 ★
Ataru — High-performance AI memory retrieval for local agent history — a Rust search core (SDK / API / JSON CLI) plus 351 ★
Yt Transcriber — CLI app- Give it a YouTube URL and you get a transcription with possible speaker identification and optional s 337 ★

README

transcript-to-knowledge

Turn any YouTube video into agent-ready research. Paste a link, get a clean markdown breakdown your AI agents can actually use.

Most people watch YouTube. This feeds it to your AI. The tool pulls a video's transcript, runs a structured LLM breakdown (and an optional deep technical dive), and writes durable markdown to disk. Point Claude, an agent, or your own RAG pipeline at the output and it absorbs the content at machine speed.

$ ttk https://youtu.be/7huCP6RkcY4 --deep
  transcribing (captions)... ok
  breakdown... ok
  deep dive... ok
-> research/cole-medin/i-built-self-evolving-claude-code-memory/
   BREAKDOWN.md  BREAKDOWN-deep.md  transcript.md  meta.json

Why

YouTube is one of the best sources for learning AI engineering, infra, CAD, and everything else moving too fast for documentation to keep up. Nobody has hours to watch all of it. This builds a research vault instead: a folder of structured notes your agents can read and implement from. It is a recursive learning pattern. The video makes you smarter once; the note makes every agent you run smarter on demand.

What you get per video

File	What's in it
`BREAKDOWN.md`	The executive summary: thesis, key concepts, architectures, golden nuggets, tools named, and how to apply it.
`BREAKDOWN-deep.md`	(with `--deep`) Transcript-grounded technical depth: exact numbers, verbatim command sequences, design rationale, quotes, and what the video leaves underspecified.
`transcript.md`	The full transcript.
`meta.json`	Video metadata + the run's provider/model/engine.

Install

**New to this and not a coder?** Open [AI-SETUP-PROMPT.md](AI-SETUP-PROMPT.md). It walks you through getting the repo onto your computer and pasting one prompt that has your AI assistant install everything, pick a provider, and run your first video for you.

If you are comfortable with a terminal, requires Python 3.10+.