Amuse is the flagship demo application for the TensorStack SDK, showcasing high-performance local AI image, video, audio and text generation through a modern, extensible .NET architecture.
- Automatic installation of an isolated, Python environment.
- Safetensors, GGUF, and ONNX support.
- Video Editor for generated or local content.
- Image/Video Upscale for static and moving media.
- Feature Extraction from images and video.
- Video Interpolation for frame rates and slow-motion.
- Image Inpaint to remove objects or fill areas.
- Advanced Image Editing with selection and masking tools.
- Voice Generation (Supertonic).
- Speech Recognition (Whisper).
- Media Gallery for organization and management.
- Lora/ControlNet Support for output control.
- Z-Image
- Qwen
- FLUX.1
- FLUX.2
- Chroma
- Kandinsky5
- StableDiffusion-XL
- StableDiffusion-3
- Ernie Image
- Anima
- JoyAI Image
- PRX-Pixel
- Krea2
- GLM Image
- Ideogram 4
- LTX
- LTX-2
- Wan 2.2
- CogVideoX
- Kandinsky5
- SkyReels-V2
- Helios
- Motif
- AnyFlow
- MiniMax-H3
- ACE-Step
- ACE-Step XL
- Whisper
- Supertonic v3
- LongCat Audio
- Qwen 3.0
- Qwen 3.5
- Qwen 3.6
- Qwen 3.8
- Gemma 4.0
Amuse provides multiple GPU backends, allowing you to choose the best option for your hardware and operating system.
NVIDIA GPUs can use CUDA 13.0 for native GPU acceleration. RTX-enabled cards are strongly recommended for the best generation performance, as they provide Tensor Cores specifically designed to accelerate AI workloads. Legacy architectures such as Pascal and Maxwell may work, but are not recommended for optimal performance.
NVIDIA driver version
580.65or later is required for CUDA 13.0 compatibility.
AMD GPUs can use ROCm 7.14.0 for native GPU acceleration. ROCm provides support for compatible Radeon and Instinct hardware, offering an alternative to CUDA for AMD-based systems. Performance and supported features may vary depending on the specific GPU architecture and operating system.
AMD driver version
26.8.1or later is required for ROCm 7.14.0 compatibility.
Vulkan provides a cross-vendor GPU backend that can be used on compatible hardware without requiring CUDA or ROCm. This makes it particularly useful for GPUs and systems that do not have access to a supported CUDA or ROCm environment.
Vulkan support depends on the capabilities of the installed GPU and Vulkan driver.
Windows 11: Users may need to
Run As Administratorduring environment creation if environment setup fails.
Note: Actual GPU compatibility and performance depend on the GPU model, driver version, operating system, and backend being used.
Amuse is made possible in part by these excellent open-source projects
PdfPighttps://github.com/UglyToad/PdfPigMarkdighttps://github.com/xoofx/markdigSeriloghttps://github.com/serilog/serilogColorCodehttps://github.com/CommunityToolkit/ColorCode-UniversalTensorStackhttps://github.com/saddam213/TensorStackDiffusershttps://github.com/huggingface/diffusersTransformershttps://github.com/huggingface/transformersStableDiffusion.cpphttps://github.com/leejet/stable-diffusion.cpp
Want to contribute to Amuse? I'd love to hear from you!
All development, source-code changes, issues, pull requests, and other contributions should be made through the main development repository:
https://github.com/saddam213/TensorStack
This repository is dedicated solely to compiled release builds of Amuse. Please use the main development repository to report issues, suggest improvements, contribute code, or get involved with the project.
Thank you to everyone who contributes to Amuse and helps make the project better!
The compiled builds distributed in this repository are covered under the Amuse 2.0 License. This licensing applies to the compiled releases due to the inclusion of bundled third-party software and their respective licenses.
If you compile Amuse yourself from the main development repository:
https://github.com/saddam213/TensorStack
the Amuse source code is licensed under the Apache License 2.0. When creating your own compiled builds, you are responsible for obtaining, including, and complying with the licenses of any third-party software and dependencies you choose to bundle.
- Compiled releases: Amuse 2.0 License
- Amuse source code: Apache License 2.0
- Self-compiled builds: You are responsible for the licensing of bundled third-party software