Back to the catalog

Cole Medin - AI Knowledge Base

Bundle OKF 0.1 · 3 conceitos · coleam00/cole-medin-knowledge-base

Open source Repository Open in the app JSON README (API)

About

# Cole Medin - AI Knowledge Base

> **Coverage: 198 long-form videos** (2023-04-05 to 2026-07-15), the full channel back catalog, compiled into **243 concept pages** and **242 entity pages** (198 tools, 13 people, 31 organizations) with **198 source summaries** over 925k words of transcript. This is an [Open Knowledge Format](https://github.com/GoogleCloudPlatform/knowledge-catalog) (OKF v0.1) bundle and a Karpathy-style LLM wiki: a synthesized, cross-linked graph of everything Cole teaches on [his channel](https://www.youtube.com/@ColeMedin), navigable with no database and no embeddings.

**If you are an agent:** read [SCHEMA.md](SCHEMA.md) for how this bundle is structured, then use this index to navigate. Follow the relative links into the pages a question needs, do not load the whole bundle. Every page ends with a `## Sources` section citing the video(s) and timestamps it came from, so you can always trace a claim back to the transcript. When you cannot answer from a page, say so r

Details

Kind
OKF bundles
Topic
AI, RAG & memory
Publisher
coleam00
Origin
okf_github
Category
dados
Version
0.1
Stars
117
Forks
38
Open pull requests
2
Last push
2026-07-28T11:11:42Z
Repository state
ativo
Language
JavaScript
Added
2026-09-09 19:04:11
Updated
2026-09-09 19:04:11
Origin id
coleam00/cole-medin-knowledge-base:index.md

README

---
type: overview
title: "Cole Medin AI Knowledge Base"
description: "An Open Knowledge Format (OKF) wiki mined from Cole Medin's entire long-form YouTube catalog, built to be dropped next to any project as agent-ready reference."
tags: [readme, overview]
updated: 2026-07-21
---

# Cole Medin - AI Knowledge Base

A synthesized, densely cross-linked knowledge base mined from [Cole Medin's](https://www.youtube.com/@ColeMedin) **entire long-form YouTube catalog** - agentic coding, AI engineering, RAG, harnesses, memory systems, and more. It is an [Open Knowledge Format](https://github.com/GoogleCloudPlatform/knowledge-catalog) (OKF v0.1) bundle and a Karpathy-style [LLM wiki](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f): plain markdown + YAML frontmatter, navigated by index and relative links, no database and no embeddings required.

**The point:** drop this folder next to your project and any coding agent can answer questions from Cole's videos with citations back to the exact video and timestamp - zero integration, no RAG stack, no API keys.

## What's inside

- **`concepts/`** - the ideas, techniques, patterns, and mental models (the PIV loop, context engineering, agentic search, and so on), each synthesizing every video that discusses it.
- **`entities/`** - the tools, people, and organizations (Claude Code, Archon, MCP, Anthropic, …).
- **`sources/`** - one summary page per video, with provenance to the raw transcript.
- **`raw/`** - the immutable, timestamped transcripts (the source of truth).
- **`index.md`** - start here (after `SCHEMA.md`). The compiled wiki *is* the search index.

## Use it with an agent (copy-paste)

Point your coding agent at this bundle with a prompt like:

```
Use the Cole Medin AI Knowledge Base as a reference:

1. Clone it next to my project (skip if I already have it):
   git clone https://github.com/coleam00/cole-medin-knowledge-base.git

2. Read `cole-medin-knowledge-base/index.md`, then `cole-medin-knowledge-base/SCHEMA.md`.
   This is an Open Knowledge Format (OKF) bundle - a linked wiki of concepts and entities
   mined from Cole Medin's videos. Navigate it the OKF way: read the index, follow the
   relative links into `concepts/`, `entities/`, and `sources/`, and open only the pages a
   question needs. Do not load the whole folder.

3. When I ask a question, answer from the knowledge base and cite the concept/entity pages
   you used and the source video(s) they came from (each page ends with a `## Sources`
   section listing the videos and timestamps). If something is not covered, tell me instead
   of guessing.
```

## Build your own (from any YouTube channel)

Three Claude Code skills are included that replicate the full pipeline - just point them at a channel. Each fetches transcripts using a different method, then walks you through the extract-canonicalize-write process described below.

| Skill | Method | API Key | Cost | Best for |
|-------|--------|---------|------|----------|
| `/channel-to-kb` | pytubefix + youtube_transcript_api | None | Free | Quick setup, local machines |
| `/channel-to-kb-ytdlp` | yt-dlp | None | Free | Most reliable, captures publish dates |
| `/channel-to-kb-supadata` | Supadata API | Required | $17+/mo | No IP issues, AI fallback for uncaptioned videos |

```bash
# Example: build a KB from 3blue1brown's channel using yt-dlp
/channel-to-kb-ytdlp @3blue1brown
```

Each skill fetches all transcripts as `raw/*.md` files, then guides Claude through the same pipeline used to build this bundle: extract concepts from each transcript, canonicalize (merge duplicates into single canonical pages), write cross-linked concept/entity/source pages, and validate with `lint.py`. The full pipeline reference is at `.claude/references/pipeline-guide.md`.

For small channels (under ~30 videos), the whole build fits in one Claude Code session. For larger channels, the skill batches the work and you can resume across sessions.

## How it was built (reproducible)

1. **Source.** Full transcripts for every long-form video are pulled from the source database and written as immutable `raw/<slug>.md` files (timestamped), plus `raw/manifest.json`.
2. **Extract.** Each transcript is mined for the concepts and entities it teaches, with timestamped quotes.
3. **Canonicalize.** All candidates are deduplicated into one taxonomy so each durable idea is a single page synthesizing every video that covers it (synthesis, not one-page-per-video).
4. **Write & link.** Concept/entity/source pages are written and cross-linked with typed relationship headings.
5. **Validate.** `python scripts/build_indexes.py` regenerates the indexes; `python lint.py` gates conformance, link integrity, index coverage, and orphans.

For the full story - the two-pass agent architecture, the tooling, and the validation results (citation integrity, adversarial QA, recall) - see [`docs/MAKING-OF.md`](docs/MAKING-OF.md) and [`tools/`](tools).

See [`SCHEMA.md`](SCHEMA.md) for the full contract and [`docs/ingestion-workflow.md`](docs/ingestion-workflow.md) for the step-by-step. Coverage and growth model are in [`roadmap.md`](roadmap.md).

## License / provenance

Knowledge is synthesized from publicly available YouTube videos by Cole Medin; each page cites its sources. Raw transcripts are included under `raw/` for provenance and auditability.

More