This project provides a specialized "skill" for AI agents, enabling them to automate video editing tasks within JianYing (the Chinese version of CapCut). It empowers agents with programmatic control over video generation, editing, and exporting, targeting users who need to integrate AI-driven video production into automated workflows. The primary benefit is the ability to create complex video content, including generative effects and web-based animations, entirely through AI commands.

How It Works

The skill acts as an intermediary, translating AI agent instructions into Python scripts that interact with the JianYing application. It leverages UI automation to control JianYing, simulating user actions for editing, applying effects, and exporting. Key innovations include a "Web-to-Video" feature that captures and imports web animations (HTML, JS, Canvas, SVG) as video assets using Playwright, and a "Semantic Footage Match" system that analyzes video content, audio, and subtitles to intelligently select and sync relevant clips from a user's library. The project also supports advanced AI-driven pipelines, integrating with other AI models for scriptwriting and asset generation.

Quick Start & Requirements

• Installation: A one-click installation script is available for Windows users via PowerShell (irm is.gd/rpb65M | iex). Manual installation involves cloning the repository into specific agent directories (e.g., .agent/skills/jianying-editor, .trae/skills/jianying-editor).
• Prerequisites:
  • JianYing Version: Crucially, JianYing Professional Edition v5.9 or lower is required. Newer versions (6.0+) are incompatible due to disruptive UI pop-ups. A download link for v5.9 is provided.
  • Python Dependencies: uiautomation, playwright, pynput, edge-tts, pymediainfo. Requires playwright install chromium.
  • AI Agent: Designed for integration with agents like Trae, Antigravity, Claude Code, and Cursor.
• Links: Installation tutorial available on Bilibili.

Highlighted Details

• Web-to-Video: Real-time capture and import of web-based animations (JS/Canvas/SVG) as video VFX assets.
• Semantic Footage Match: AI-driven selection of video clips based on semantic analysis of dialogue and subtitles.
• End-to-End Automation: Supports full pipelines from AI scriptwriting and asset generation (e.g., Imagen 3, Grok 3) to JianYing project assembly and automated export.
• Screen Recording Enhancements: Features "Smart Zoom," automatic red circles around the cursor, and smooth mouse following for clearer tutorials.

Maintenance & Community

The project appears actively maintained, with updates noted up to February 9, 2026 (v1.4). Updates are managed via git pull. No specific community channels (like Discord or Slack) or notable contributors/sponsorships are mentioned in the README.

Licensing & Compatibility

The license type and terms are not specified in the provided README text. Compatibility is heavily restricted by the mandatory requirement for JianYing v5.9 or lower. While manual installation is possible on Mac/Linux, the reliance on UI automation and specific dependencies may favor Windows environments.

Limitations & Caveats

The most significant limitation is the strict dependency on JianYing v5.9 or earlier; any update to JianYing risks breaking the automation scripts. Automated export requires the user to avoid interacting with the mouse and keyboard during the process, as it relies on simulated inputs. The absence of a specified license raises concerns regarding usage rights and commercial compatibility.