feat: AI-generated player content + regrouped Video Player options

[eitanp461]

Summary

Two changes to the Widget → Video Player operation:

1. New AI-Generated Content options — drive the player's on-demand AI generation, so users get captions, transcripts, title/description, and chapters with no files to author.
2. Regrouped the Video Player UI — the 19 controls are no longer a flat wall; "Advanced Player Options" is gone, replaced by purpose-named collapsible groups with the source essentials kept top-level.

AI-Generated Content

A new collapsible section on the node:

• Generate Captions — auto-generates a transcript shown as a toggleable captions track, in the video's original spoken language.
• Subtitle Languages — optional, comma-separated target language codes for AI-translated subtitle tracks. Empty by default.
• Generate Title / Description / Chapters — AI-generated metadata + chapter markers. The chapters button is grouped under (and gated on) Generate Chapters, so it can never surface a button with no chapters source.

These are emitted into player_config (the JS player.source() config), not the iframe embed URL — matching how the player exposes them.

Out-of-the-box safety

• The collection defaults to {} — with nothing added, the node emits zero AI keys and behaves exactly as before. Fully opt-in.
• Captions/title/description/chapters are unsigned actions enabled by default on most accounts → work with no setup.
• The only add-on-gated feature (translated subtitles) is never triggered by default: Subtitle Languages is empty, so a user must deliberately enter codes. The README documents a Google Translate add-on setup CTA for when they do.

New Video Player layout

VIDEO PLAYER
├─ Public ID · Type · Transformation · Poster · Source Types   (top-level essentials)
├─ ▸ Playback           Autoplay · Muted · Loop · Plays Inline · Show Controls · Big Play Button
├─ ▸ Size & Layout      Fluid · Width · Height · Aspect Ratio · Crop Mode
├─ ▸ AI-Generated Content   Captions (+Label, Languages) · Title · Description · Chapters (+Button)
├─ ▸ Appearance         Skin · Base/Accent/Text Color · Font Face
└─ ▸ Player Features     Floating · HDR · Picture-in-Picture · Seek Thumbnails

Also in this PR:

• Poster promoted to a top-level source essential (it's about the source, like Transformation — not playback or chrome).
• Source Types description rewritten to lead with plain meaning + an HLS/MP4 fallback tip, keeping the format-pin caveat. Name kept to mirror the Cloudinary API's sourceTypes property.
• README documents the AI options + the Google Translate add-on enable steps.

Backward compatibility

⚠️ This relies on 0.0.9 being the currently active (latest published) npm version.

The only version ever published to npm is 0.0.9, which has no Video Player operation at all. Regrouping these fields into collections renames their parameter keys — but since no released version ever exposed Video Player fields, no saved workflow can reference them, so nothing breaks.

The backward-compatibility-check reports 0 breaking changes against the 0.0.9 baseline; every new/moved param registers as additive.

This holds only while 0.0.9 is the latest published version. Once a release that ships the flat Video Player fields goes out, any further regrouping would require a typeVersion bump.

Bumps version to 0.2.0 (additive feature).

Testing

• npm test — 234 tests pass (widget suite extended for the new AI fields, the nested-collection reads, and the chapters-button gating).
• tsc --noEmit, eslint, npm run build — clean.
• npm run backward-compatibility-check — 0 breaking changes.

🤖 Generated with Claude Code