YouTube Lesson Caption Knowledge-Point Still Frames

Problem

When teachers need to turn selected YouTube lesson lines into courseware-friendly still frames, this skill works from the supplied subtitle text rather than inventing content. It targets 4 to 8 printed knowledge-point frames grounded in chosen lines, not re-scripting, thumbnails, or enterprise training cards.

How It Works

• Input scope: confirm the lesson, subtitle source, and card count before doing paid work.
• Slot table: record subtitle line, knowledge line, layout, canvas, and optional brand-image reference; this is a free planning artifact, not approval to generate.
• Lookup and generation: if a lookup is requested, present a six-field card; then read the current text_to_image price and call beatra.images.generate per slot.
• Remote calls: use the bundled scripts/mcp_client.py only, without REST/OpenAPI fallback.

Boundaries

It does not publish YouTube videos or invent missing subtitles, titles, or view counts. Pasted subtitles are preferred; lookup and generation are separate billing stages, and insufficient balance does not trigger automatic retries. Small rendered text is treated as a review item, and delivery relies on actual bytes, MIME, aspect, size, and task billing.