VideoJaNai (previously called AnimeJaNaiConverterGui) provides a Windows and Linux GUI for upscaling videos with ONNX models, including temporal (multi-frame) models, and frame interpolation with RIFE. Upscaling runs entirely on the GPU, with NVIDIA TensorRT or, on Windows, DirectML for AMD and Intel GPUs, through the AnimeJaNai inference engine, the same engine that powers mpv-AnimeJaNai.
On Windows, NVIDIA GeForce RTX 20 / GTX 16 series and newer GPUs upscale with TensorRT-RTX by default, which builds its engines in seconds instead of minutes. Its component (about 230 MB installed) is installed automatically for those GPUs; on other NVIDIA GPUs it's optional, from App Settings > Components. TensorRT stays installed alongside it: TensorRT-RTX output is equivalent to TensorRT's but not bit-identical, and if it can't start or can't build an engine, VideoJaNai falls back to TensorRT automatically and shows a warning in the status bar (and in the preview). To choose a backend yourself, pick TensorRT, TensorRT-RTX or DirectML under Backend in App Settings.
Download either the AppImage or the linux-x64.tar.zst from the latest release. An NVIDIA GPU and driver are required.
- AppImage: make it executable and run it. Run it once with
--installto add a menu entry. The TensorRT kernels for your GPU and the RIFE models download on first launch into~/.local/share/VideoJaNai. - Tarball: extract it anywhere and run
./install.shfrom inside the extracted folder. This adds a menu entry and installs the components for your GPU. The folder stays in place and updates itself in-app.
- OpenModelDB: Repository of AI upscaling models.
- chaiNNer: General purpose tool for AI upscaling, which can be used to convert PyTorch (*.pth) models to ONNX models.
- mpv-upscale-2x_animejanai: A video player which supports real-time upscaling using compact ONNX models on higher end GPUs.
- traiNNer-redux: Software for training upscaling models
- MPV_lazy
- vs-mlrt
- vs_temporalfix: the temporal fix models (Apache-2.0)