hao0321/video-autopilot-kit

Video Autopilot Kit is a Python-based framework for automating the production of YouTube videos and short-form content. Rather than providing a rigid, one-size-fits-all workflow, it works as a configurable template: you answer a series of questions about your channel's brand, voice, and content strategy, and the setup process turns those answers into a production system tailored to your specific content. The primary editing path relies on CapCut Desktop, a free video editing application. Because CapCut has no public programming interface, automation is achieved by having an AI assistant (such as Claude with Computer Use enabled) literally operate the CapCut application on-screen, clicking buttons, applying templates, and exporting files. Computer Use support in your AI tool is a hard requirement for the main workflow to function. A secondary path uses only ffmpeg, a command-line video processing tool, for silent vlogs that have no spoken narration. Beyond the CapCut automation helpers, the toolkit includes utilities for post-export ffmpeg processing such as looping background music, trimming to voice length, and re-encoding for platform compatibility. It also provides AI subtitle correction and b-roll placement auditing. The templates directory contains blank fill-in files for voice, branding, algorithm strategy, and community settings. All configuration examples and templates ship empty so no personal data from the original author is included. The README is written in Traditional Chinese. The project is MIT licensed and described by its author as extracted from a working personal creator system, with the structure made open for others to adapt to their own channels.