Auto JianYing Video Editing with Voiceover and Subtitles

Problem

Turning a mix of PPT slides, images, text documents, and video clips into a microlesson or short promotional video often fails at integration: segment order, narration alignment, subtitle timing, JianYing draft compatibility, and MP4 export are tightly coupled. Different JianYing Pro versions can also break draft schemas, making projects hard to open or edit.

How it works

The skill processes assets as a pipeline. It scans PPT, txt/md/docx, images, and video, identifies valid pages and segment order, and renders PPT pages to 1920x1080 PNG files when PowerPoint is available. It then applies per-segment Chinese narration, generates audio with edge-tts, composes each visual and audio pair into a clip, and builds an SRT timeline from cumulative segment durations.

Draft generation is version-aware. It detects the local JianYing version and writes a v10.9 / v11.2 compatible project from a real v11.2 draft skeleton; older versions can upgrade it on open. The main MP4 deliverable is produced by ffmpeg, which concatenates clips, burns subtitles, and outputs H.264+AAC without calling JianYing export. This keeps MP4 generation stable across versions, even when JianYing is absent.

Boundaries

It fits teaching, lecture capture, and image/text-to-video workflows where both a playable video and an editable project are useful. If assets do not include PPT, Office rendering can be skipped; if only the final video is needed, draft generation can be ignored. Note that PowerPoint COM requires a visible window, ffmpeg SRT paths should be relative, and JianYing GUI export still requires manual action.