The content model

A Vidext module is typed, validated data — not opaque media. Understanding the shape pays off whenever you build, edit, or interpret modules programmatically. (For the product-level view, start with Scenes and layouts.)

The snapshot envelope

A module's renderable content is a module snapshot: a versioned JSON envelope (schema version, creation time, the use case it was planned with) wrapping the module itself:

ModuleSnapshot
└── Module { id, title, description? }
    └── Chapter[] { id, title, description? }
        └── Scene[]

Snapshots are validated on read and write — a malformed snapshot is rejected, never half-served.

Scenes: a discriminated union

Each scene is:

Scene { id, layoutId, background?, content, audioNarration?, sourceRefs? }

layoutId is the discriminant: it selects which content schema applies. The ten layout ids:

start · chapter-intro · bullet-points · bullet-media-points · stack-cards
content-avatar · image-avatar · video · quizzes · outro

Each layout also carries metadata the system plans with — its teaching role (opening / core / closing / assessment), its presentation family, and its image requirements.

sourceRefs? is scene-level provenance: an optional array of references the builder records at write time, each pointing to the evidence that grounded the scene — a documentId (an uploaded source) or a url (a web finding), plus an excerpt and a relevance note. It survives through publish, so built modules carry traceability back to their sources.

Components: the shared atoms

Layout content composes reusable component schemas — headers, bodies, bullets, cards, video players — plus two worth special attention:

• Narration — spoken text with a reveal duration. Narration is both the teaching script and the timing source for audio, subtitles, and auto-advance. A separate hidden audio-narration track (segments anchored to content) drives generated speech without altering visible copy.
• Quizzes — quiz-select (3–4 answers, exactly one correct, with an explanation) and quiz-boolean (true/false with explanation). A quiz scene holds one to three of them.

Draft vs publishable: two strictness levels

Every schema has two variants:

• Draft — permissive: fields may be empty mid-generation.
• Publishable — strict: non-empty strings, positive narration durations, at least one chapter and scene.

Generated content must pass the publishable contract before it can be published. This is the type-level expression of the product's draft and published gate.