Ask Hormozi

Ask Hormozi ships a Markdown corpus for Alex Hormozi's MoreMozi channel and gives AI coding agents a skill for answering questions with video-level, timestamped sources. Installation builds only the local QMD index; it does not download video, audio, captions, or transcripts.

The current corpus snapshot contains 2,039 videos, 2,039 readable transcripts, and 9,330 timestamped segments.

What it does

• Ships a catalog of every MoreMozi video with its YouTube URL.
• Ships a readable Markdown transcript for each video.
• Ships 90-second Markdown segments with video metadata and timestamps.
• Registers those segments as a dedicated QMD collection.
• Installs an ask-hormozi skill for Codex and Claude Code.
• Returns retrieved context with clickable YouTube timestamps so the agent can answer without inventing attribution.
• Syncs incrementally: reruns skip videos already transcribed.

Install as an HQ pack

Install from the HQ marketplace:

hq install marketplace:ask-hormozi
bash core/scripts/hq-ask-hormozi setup

Or install the full source package directly from GitHub:

hq install https://github.com/poseljacob/ask-hormozi
bash core/scripts/hq-ask-hormozi setup

The second command prepares the Markdown already bundled in the pack and creates the local QMD index. Marketplace installs unpack a compressed copy of the same timestamped Markdown locally to stay within the marketplace artifact limit. Neither path downloads video, audio, captions, or transcripts during setup.

Start a fresh Codex or Claude Code session from HQ, invoke /ask-hormozi, and ask a normal question:

What does Hormozi recommend for improving an offer?

The skill checks the index on each use and creates it automatically when it is missing. Direct HQ runtime commands are also available:

bash core/scripts/hq-ask-hormozi search \
  "How should I price and position my offer?"
bash core/scripts/hq-ask-hormozi doctor
bash core/scripts/hq-ask-hormozi refresh

Standalone install

Requirements: macOS or Linux, Python 3.10+, and curl. The setup script installs the package in an isolated virtual environment and installs QMD from its official installer when QMD is not already available.

git clone https://github.com/poseljacob/ask-hormozi.git
cd ask-hormozi
./setup.sh

Ensure ~/.local/bin is on your PATH, then search directly:

ask-hormozi search "How should I price and position my offer?"

Or invoke the installed /ask-hormozi skill in Codex or Claude Code and ask a normal question:

What does Hormozi recommend for improving an offer?

HQ pack contents

The root package.yaml declares two contributions:

• skills/ask-hormozi — the portable Codex/Claude retrieval skill.
• scripts/hq-ask-hormozi — the HQ runtime that uses the bundled Python module and corpus without a separate package install.

HQ wires those contributions to .claude/skills/ask-hormozi and core/scripts/hq-ask-hormozi. The corpus remains inside core/packages/hq-pack-ask-hormozi/corpus.

Commands

ask-hormozi sync [--limit N] [--jobs N] [--force]
ask-hormozi catalog [--output-dir corpus]
ask-hormozi audit-captions [--jobs N]
ask-hormozi index
ask-hormozi configure --data-dir PATH
ask-hormozi search "question" [--limit N] [--format markdown|json]
ask-hormozi doctor

corpus/
├── catalog.json
├── caption-coverage.json
├── transcript-manifest.json
├── episodes/
│   └── VIDEO_ID.md
├── metadata/
│   └── VIDEO_ID.json
├── transcripts/
│   └── VIDEO_ID.md
└── segments/
    └── VIDEO_ID/
        ├── 000000.md
        ├── 000090.md
        └── ...

setup.sh configures the CLI to use the checked-out corpus/ directory. index owns the QMD collection named ask-hormozi; it replaces that collection registration without running a global QMD update.

Refreshing

Update the HQ pack and re-index:

bash core/scripts/hq-ask-hormozi refresh

For a standalone checkout, pull the latest prebuilt corpus and re-index:

git pull
./setup.sh

End-user installation never downloads from YouTube. The maintainer-only ask-hormozi sync command rebuilds corpus source files for an authorized release.

Maintainers can refresh non-transcript release metadata independently:

ask-hormozi catalog --output-dir corpus
ask-hormozi audit-captions --output-file corpus/caption-coverage.json

The caption audit resumes from its existing output. If YouTube throttles a large pass, retry only failed videos conservatively:

ask-hormozi audit-captions --jobs 1 --delay 5

The maintainer commands use yt-dlp's unauthenticated web_embedded YouTube client, falling back to its tv client when a video disables embedding. They do not read browser cookies or require a YouTube account.

Retrieval and citations

The skill runs QMD BM25 search through the package CLI. Each result is enriched from the matching local segment:

{
  "title": "Video title",
  "start_seconds": 540,
  "context": "Retrieved transcript passage...",
  "citation_url": "https://www.youtube.com/watch?v=VIDEO_ID&t=540s",
  "transcript_source": "automatic_captions"
}

The skill must paraphrase by default, distinguish Hormozi's statements from its own synthesis, and cite material claims as:

[Video title — 09:00](https://www.youtube.com/watch?v=VIDEO_ID&t=540s)

Development

python3 -m venv .venv
.venv/bin/python -m pip install -e .
.venv/bin/python -m unittest discover -s tests -v

Bundled-corpus smoke test:

.venv/bin/ask-hormozi configure --data-dir "$PWD/corpus"
.venv/bin/ask-hormozi index --data-dir "$PWD/corpus"
.venv/bin/ask-hormozi search "pricing value equation irresistible offer"

Content and affiliation

This project is not affiliated with or endorsed by Alex Hormozi, Acquisition.com, MoreMozi, or YouTube.

The transcript corpus is redistributed with permission. Copyright in the video-derived corpus remains with its respective owner and is not relicensed under the repository's or package manifest's MIT code license. See corpus/NOTICE.md. No video or audio binaries are included. Verify important details against the linked video because captions can contain errors.