AI JianYing Draft Generator

Problem

Manually turning scripts, transcripts, and selected clips into JianYing drafts often leads to mismatched timestamps, overlapping segments, narration covering original audio, and inconsistent output paths. This skill is aimed at commentary-style videos where transcript/scene analysis or storyboard JSON already exists, and it makes the “storyboard to inspectable draft” path explicit.

How It Works

• Drafts are written in the user’s project directory, not the Skill directory; unique draft names are used by default, and --overwrite requires explicit approval.
• Narrated, original-audio, subtitle, transition, and audio rules are handled through the workflow and rules/ files.
• storyboard_validator.py checks timestamps, overlaps, durations, and schema; jy_wrapper.py uses the draft API; draft_inspector.py inspects the generated draft.
• mode: "original" preserves source audio; mode: "narration" adds explanatory lines. If TTS exceeds the planned duration, it warns and rewrites or extends instead of silently overlapping scenes.

Boundaries

Best for commentary, long-video summaries, and storyboard-to-draft tasks. It is not suitable when there is no transcript, no scene analysis, or timestamps cannot be confirmed. Dynamic cloud resolution is disabled by default, cloud downloads require explicit approval, and exports must be verified by inspecting the draft or resulting file before claiming success.