Demo Material

HVE Core publishes a training deck, a narrated video, and a browser slide deck for each of four depth levels. Each level targets a different audience and builds on the one before it.

Demo Material

Level	Audience	Length	Watch	Present	Download
L100	New contributors and first-time HVE Core users	4 to 6 min	L100 video and transcript	L100 slides	L100 deck, L100 MP4
L200	Contributors ready to follow a guided task	6 to 8 min	L200 video and transcript	L200 slides	L200 deck, L200 MP4
L300	Engineers choosing an applied workflow	8 to 10 min	L300 video and transcript	L300 slides	L300 deck, L300 MP4
L400	Maintainers and contributors extending HVE Core	10 to 12 min	L400 video and transcript	L400 slides	L400 deck, L400 MP4

L100 and L200 videos show the deck slides. L300 and L400 add live VS Code captures of the repository files they discuss.

The browser slides use the same presenter controls as the HVE Core presentations: arrow keys to move, a slide index, speaker notes, the level's sources, and a reading view. Each is one self-contained HTML file, so you can download it and present offline.

Accessibility

Every level is published with:

• Captions embedded in the MP4 and as a separate WebVTT file, generated from the exact narration text
• A video page with a captioned player and a full transcript listing each slide's title, on-screen text, image descriptions, and narration
• Narration that voices what each slide shows and describes each live capture, so the audio carries the visual content
• A deck with a title on every slide, alternative text on every image, and the document language set, so screen readers can navigate it
• Browser slides with keyboard navigation, labelled slides, a reading view that reflows on small screens, and speaker notes available to every viewer
• Text colors that meet the WCAG 2.2 AA contrast minimum of 4.5:1

The render workflow checks each item it can verify and withholds a level that fails. Caption timing within a slide is estimated from sentence length, so captions can lead or trail the voice by a moment. Whether the narration fully describes each visual cannot be checked by a machine and needs human review.

How the Material Stays Current

The material is rebuilt by two scheduled workflows, and only for the levels whose source documents changed:

1. The Demo Material Author agentic workflow runs weekly. A deterministic step compares each level's pinned sources with the commit its current material was built from, and the agent writes new slide content only for the levels that changed. When nothing changed, the agent does not run.
2. The Demo Material Render workflow builds each authored level into a deck, narration, a video, and browser slides from the same slide content, then scores the checks a machine can verify: narration paired to every slide, the video landing inside the level's length, readable live captures, the pinned house style, the accessibility items above, and browser slides that open offline with no slide overflowing. A level that fails any check keeps its previous files.
3. The documentation deployment publishes the latest passing files on this page.

The narration uses an offline Piper voice (en_US-joe-medium, CC0), so the builds need no cloud speech service. The level contracts and sources define what each level covers.

Build Status

The build index records, for each level, the commit it was built from, when it was rendered, its measured length, and the result of every check, including the most recent failed attempt.

NOTE

The decks, videos, and slides are drafted by an AI agent from this repository's documentation and checked automatically, not reviewed by a person before publication. Review a deck before presenting it to an external audience. A level's links resolve after its first successful render.

🤖 Crafted with precision by ✨Copilot following brilliant human instruction, then carefully refined by our team of discerning human reviewers.