Lingji Cut (灵剪) is an open-source, local-first AI video creation workbench designed for content creators. It integrates a comprehensive workflow, encompassing scriptwriting, asset management, AI-powered review, text-to-speech (TTS), subtitle processing, timeline editing, and video export into a unified desktop environment. This tool aims to streamline the entire video production process, allowing users to generate polished videos from raw materials or existing audio and subtitles.

How It Works

The core of Lingji Cut is its integrated, local-first approach. It manages content creation stages within a desktop application, starting with an AI-assisted writing workbench for script development (.md files) featuring version history and AI review capabilities. The workflow extends to a unified video workbench with a professional timeline editor supporting multiple audio, video, subtitle, and AI-generated visual card tracks. It leverages a multi-provider AI configuration, allowing users to connect various LLMs (OpenAI-compatible, Gemini, LM Studio), image generation services, and TTS providers (like MiniMax). Additionally, it integrates with Claude ACP Runtime and MCP pipelines for external agent-based automation and control.

Quick Start & Requirements

• Install: npm install (uses npm mirrors by default; may require setting ELECTRON_MIRROR and npm_config_disturl environment variables for specific network environments).
• Development: npm run dev
• Build: npm run build
• Package: npm run dist:mac (macOS) or npm run dist:win (Windows).
• Test: npm test
• Prerequisites: Node.js, npm, Electron, React, TypeScript, Remotion. Specific npm mirror configurations are recommended for certain network conditions.
• Links: The README serves as the primary documentation.

Highlighted Details

• AI Writing Workbench: Manages .md scripts with multi-file tabs, resource management, search/replace, version history, AI generation, and AI review.
• One-stop Video Workbench: Unified interface for asset management, preview, inspector, timeline, and export configurations.
• Automated Dubbing Workflow: Triggers TTS, subtitle parsing, content analysis, cover candidate generation, and visual card arrangement from scripts.
• Multi-Provider AI Configuration: Supports OpenAI-compatible models, Gemini, LM Studio, and various image/TTS providers.
• Agent/MCP Integration: Connects to Claude ACP Runtime and provides MCP tools for external agents to control project creation and task orchestration.
• Remotion Export: Renders compositions using Remotion for H.264 MP4 output with configurable resolution and quality.
• Local-First Architecture: Project files are stored locally; no real API keys are saved in the repository.

Maintenance & Community

• Contact: X/Twitter: @LYoqu60097, WeChat: yoqu2020.
• Community: No explicit links to Discord/Slack or forums are provided.

Licensing & Compatibility

• License: Apache License 2.0.
• Compatibility: The Apache 2.0 license is permissive and generally compatible with commercial use and linking in closed-source projects.

Limitations & Caveats

The application is desktop-first with a minimum window size constraint (~1100x760) and is not optimized for mobile. The script-to-video workflow relies on MiniMax TTS, and Douyin (TikTok) import depends on the external bcut transcription service. AI analysis, generation, TTS, and Agent Runtime functionalities require user-configured external services. macOS packaging is not yet formally signed or notarized for distribution.