About this skill

Problem

Short-video production often stalls when images, narration, audio, subtitles, and FFmpeg parameters are handled separately. Common failures include narration mismatch, audio loss during concat copy due to inconsistent stream structure, Chinese title rendering issues with drawtext, and manual chart screenshots from Excel/CSV data. This skill turns these steps into a reusable scripted workflow for stable MP4 generation.

Core Capabilities and Workflow

Audio-visual alignment: Generates narration for each image that describes that image's visual content, then renders audio with edge-tts.
Stable compositing: Prefers encoding each image-plus-audio segment into a complete MP4 before concatenation, preserving h264 video and aac audio streams.
Title pages: Uses PIL to generate Chinese title images and avoids FFmpeg text-rendering issues.
Data visualization: Uses generate_chart_video.py to turn Excel/CSV data into animated bar, line, pie, or bar-race charts.
End-to-end generation: Uses generate_full_video.py to move from topic to script, storyboard, images, audio, and final video.
Error handling: Uses error_handler.py for structured error codes, retries, temp-file cleanup, and logging to output/error.log.

Boundaries

Best suited for ordered image sequences, verifiable narration, and quick explanatory or chart videos. For complex images or highly creative narration, scripts still need manual review. If ffmpeg, edge-tts, or file paths are missing, run environment checks and file validation first.

Screenshots

Use Cases

Turn an ordered set of product explanation images into an MP4 with Chinese narration for internal training.
Generate an animated bar-chart video from quarterly sales spreadsheet data for a business review.
Produce a full MP4 from a topic by running the end-to-end script, storyboard, image, and audio pipeline.
Diagnose FFmpeg or edge-tts failures during video generation and produce a concrete troubleshooting report.

Best For

Instructors who need to turn ordered slide-like images into narrated training videos.
Analysts who need to convert spreadsheet data into animated chart videos for reviews.
Content operators who need topic-driven short-video scripts and draft MP4 outputs.
Engineers who need to debug FFmpeg and edge-tts dependencies in the video pipeline.