Wunjo Make

[Documentation](https://docs.wunjo.online)
  

[Issue](https://github.com/wladradchenko/wunjo.wladradchenko.ru/issues)
·
[Discussions](https://github.com/wladradchenko/wunjo.wladradchenko.ru/discussions)
·
[Tutorial](https://youtube.com/playlist?list=PLJG0sD6007zFJyV78mkU-KW2UxbirgTGr&feature=shared)

Versions

The main branch contains Wunjo Make v3, a free and open-source video editor. Features such as video generation, face swap and lip sync now come as plugins that you install only when you need them: Video Generation, Depth Map, Face Swap and Live Portrait (Face Toolkit), Lip Sync, Voice Conversion and Audio Separator (Voice Toolkit) and an MCP agent. Plugins are available at wunjo.online/product/plugins, and the documentation lives at docs.wunjo.online.

Wunjo CE v2 and Pro v2 has been discontinued. Wunjo Make is now the single edition of Wunjo, free for everyone.

This is a big change from v2, and it was made on purpose. Moving these tools into plugins keeps the editor light, lets each model be updated on its own schedule, and makes it possible for anyone to add new tools. And it opens up the possibility of macOS support.

Looking for Wunjo v2 Community Edition?

Wunjo v2 CE is not a full video editor. It is a standalone toolkit built on neural networks, with Face Swap, Lip Sync, Live Portrait, removal of objects, text and backgrounds, Restyling, Audio Separator, Voice Cloning, video and image generation and Highlight Moments.

Previous versions of the product are available. If you rely on it, you can still use it.

Ready-made builds of older versions are also available at wunjo.online/product/downloads. The v2 branches stay available, but new features are developed only in v3.

What comes next

Over time, v3 will receive everything v2 could do and more, as plugins. You can also build your own plugins and share them with the community. If a v2 feature matters to you and is not in v3 yet, open an issue so we know what to bring over first.

About

Wunjo harnesses the power of advanced neural networks to transform video, image, and audio content in unprecedented ways. It allows you to seamlessly replace faces, control facial motion, and enhance lip-sync animation, while also generating videos from text, photos, or multiple images. Intelligent video editing automates cropping to any aspect ratio, selects the most interesting moments, and enables text-based montage. Wunjo can remove objects, improve video quality, restyle images and videos, clone speech, and even separate music from spoken words, giving you full creative control over every element of your media. Official website wunjo.online.

The program is a full-fledged video editor based on the Kdenlive by KDE License GPL v3 codebase. It ships an MCP server, so the agent you already use — Claude Code, Cursor, Codex — drives the editor directly: the Chat panel writes a folder with the tool registration and the instructions, you open it in a terminal, and what the agent does appears in the panel while you watch the timeline. It also ships a library of editing skills — how to cut spoken footage, time subtitles, grade a sequence, direct narration — that the agent reads before it starts, so the result is an edit rather than clips laid end to end; you can edit any of them or add your own, and a loop turns a whole scenario into steps the agent follows.

MCP

Driving the editor needs no model on your machine. If you want one anyway, a local model is a plugin you add: Qwen 3.5 is published as one, and any other can be packaged the same way. One plugin, one model — install several and pick between them in the Chat panel, or leave it empty and work from your own agent.

The new version incorporates the concept of Python plugins. You can freely create plugins using vibecoding and use them within the program for working with faces, video, and audio or content generation local or remotly by API.

The program runs on Linux, Windows, and MacOS. The MCP server itself is small and needs nothing but Python; a local model plugin is what asks for the hardware — 8GB of video memory and CUDA 12.x. Otherwise, consider using MCP via Claude Code or Cursor and plugins with API services.

Try it now

Please note

The previous version of Wunjo Make v2 is still available in the repository. It offers more functionality for local editing and content generation. The new ideology of Wunjo Make v3 is being developed with cross-platform compatibility (running on MacOS) and the ability to run on devices without a GPU and CUDA, taking into account the MSP and API for content generation and self plugin creation by neural coding.

Setup

Easy installation

For development

Linux

The application is built and run as a flatpak. Install the tooling and add Flathub once:

sudo apt install flatpak flatpak-builder
flatpak remote-add --if-not-exists flathub https://flathub.org/repo/flathub.flatpakrepo

Then build and install it. Folder build-flatpak is a scratch directory the builder creates itself, nothing has to exist there beforehand:

git clone https://github.com/wladradchenko/wunjo.wladradchenko.ru.git
cd wunjo.wladradchenko.ru/portable

# MLT pulls one of its dependencies over file://, which git refuses by default
GIT_CONFIG_COUNT=1 GIT_CONFIG_KEY_0=protocol.file.allow GIT_CONFIG_VALUE_0=always flatpak-builder --user --install --force-clean --install-deps-from=flathub --default-branch=master build-flatpak .flatpak-manifest.json

flatpak run online.wunjo.make

The KDE runtime and SDK come from Flathub on the first run of that command; MLT, FFmpeg and the other pinned dependencies are built from source, so expect the first build to take a while and some 30 GB of scratch space. Later builds only recompile the editor. Delete build-flatpak/ and .flatpak-builder/ when you are done — they are the scratch space, not the program.

To hand the result to somebody else as a single file, export it through a local repository repo is created by the command as well:

flatpak-builder --repo=repo --force-clean build-flatpak .flatpak-manifest.json
flatpak build-bundle repo wunjo-make.flatpak online.wunjo.make master --runtime-repo=https://flathub.org/repo/flathub.flatpakrepo

Windows and macOS

Both are built with KDE Craft from desktop/packaging/craft — the same source tree, no second implementation. Installers are published on GitHub Releases as they are signed and tested.

Requirements

A graphics card is optional for editing and recommended for the AI plugins. Plugins run on the processor without one, several times slower. Each plugin says what it needs in its own settings page, and downloads it there.

Plugins

Plugins are installed from a .wmplugin archive; none of them is required for the editor or for driving it from an agent. The following are published for it:

In order to install the plugin Settings > Configure Wunjo Make > Plugins > From archive and select .wmproj file. Next, you will need to click the import plugin button, and then, in the plugin tab, click the install button and download the models, if necessary.

Built-in plugin already have their own plugin tabs. However, you'll need to click the install button and then click the download button for the selected model.

For developers

In the plugins directory, you'll find a README file with information for vibecoding, how to quickly create plugins, connect various services to them, and compile them.

Updates

Update 3.0.0

• [x] Rebuilt as a desktop nonlinear editor on the MLT framework, replacing the web interface.
• [x] Multi-track timeline with transitions, effects, keyframes and time remapping.
• [x] Colour correction, audio post-production, subtitles and titling.
• [x] Rendering to practically any format through FFmpeg.
• [x] Built-in face detection for tracking or hiding a face on a clip.
• [x] Plugin platform as isolated Python environments.
• [x] Plugins for speech recognition, voice cloning, voice separation and content cutting.
• [x] Chat assistant that drives the editor through its scripting interface.
• [x] Update notifications for installations made from a file.

All information about changes and new features for Wunjo Pro can be found at CHANGELOG.

Support the Project

You can support the author of the project in the development of his creative ideas on Boosty or CloudTips. Thanks to Boosty, you will have the opportunity to use your voice to choose which functions from the Professional version will appear in the Community Edition in new updates. There are other ways to support the development of the project, more details on page.

Boosty

CloudTips

Sponsors

This open-source project is part of the SignPath OSS program and we are grateful for their support in helping us maintain and secure the project. Free code singing o Windows provided by SignPath.io, certificate by SignPath Foundation. Signed installer coming soon in releases.

Supporters and Donors

I extend heartfelt gratitude to the following individuals who have generously supported this project through donations:

• Monogo - 110$.
• S. - 50$.
• Konstantin Kravtsov - 35$.
• LEO - 5$.
• Several contributors who have chosen to remain anonymous or opted not to be listed publicly. Your support is immensely appreciated.

I sincerely appreciate the generosity of all project supporters. Your contributions enable me to continue improving and maintaining this project.

You can support the project by making a donation or subscribing for updates, news, voting, and early access to Wunjo Community Edition on Boosty or Patreon.

Contact

Owner: Wladislav Radchenko

Email: i@wladradchenko.ru

Project: https://github.com/wladradchenko/wunjo.wladradchenko.ru

Web site: wunjo.online

Premise

Wunjo comes from the ancient runic alphabet and represents joy and contentment, which could tie into the idea of using the application to create engaging and expressive speech. Vunyo (ᚹ) is the eighth rune of the Elder and Anglo-Saxon Futhark. Prior to the introduction of the letter W into the Latin alphabet, the letter Ƿynn (Ƿƿ) was used instead in English, derived from this rune.

Credits

Wunjo is built upon the remarkable work of various open-source projects. Each integrated component reflects a commitment to improving and adapting existing technologies within the collaborative landscape of open-source development. The list below highlights the projects that have been adapted and enhanced for inclusion in Wunjo:

• User Interface & Packaging: Implementations of KDE Kdenlive, Flatpak MLT framework, Qt, KDE Frameworks, FFmpeg and frei0r.
• Facial Animation & Enhancement: Adapted versions of Face Utils, LivePortrait, Insightface which provided a well-made library and some models for face.
• Image & Video Enhancement: Adapted Real-ESRGAN for superior quality enhancements
• Video Segmentation: Adaptations of Segment Anything 2
• Speech: Whisper for recognition, Silero VAD for finding where speech is

I extend my deepest gratitude to the original contributors of these technologies. Their groundbreaking work has been instrumental in advancing the capabilities of Wunjo Make. For the persistent storage and versioning of the models I have personally trained, I utilize the Hugging Face Model Storage. If you're interested in contributing to Wunjo Make, especially in the area of voice cloning for new languages, please feel free to propose your models or reach out for collaboration via GitHub Discussions or the Hugging Face.

[to top]