About this project

Wunjo Make is a free and open-source desktop video editor built on the Kdenlive codebase and the MLT framework. The main branch contains version 3, which keeps the core editor lightweight and moves neural-network features into optional Python plugins. Published plugins include Face Toolkit for face swap and Live Portrait, Lip Sync using LatentSync, Voice Toolkit for voice cloning and voice/music separation, and Cut Finder for detecting scene changes and speech boundaries. Built-in plugins provide Whisper transcription and SAM object masking. The editor offers a multi-track timeline with transitions, effects, keyframes, time remapping, colour correction, audio post-production, subtitles, and rendering through FFmpeg. It also ships an MCP server so external agents such as Claude Code, Cursor, or Codex can drive the editor through a scripting interface, with a library of editing skills available to guide the agent. A local model plugin can be installed for on-device agent use, requiring about 8GB of video memory and CUDA 12.x, while the editor itself can run without a GPU. Wunjo Make runs on Linux, Windows, and macOS. Linux builds use Flatpak, while Windows and macOS builds are produced with KDE Craft. Older v2 Community Edition releases remain available in separate branches for users who need the standalone neural-network toolkit, but new development focuses on v3. The project is licensed under GPL v3 and credits numerous open-source components including Kdenlive, MLT, Qt, FFmpeg, LivePortrait, Insightface, Real-ESRGAN, Segment Anything 2, Whisper, and Silero VAD.