SD.Next
A web interface and server for image and video generation that runs on Nvidia, AMD, Intel, Apple and CPU-only hardware.
One web UI for dozens of models, with backends for Nvidia, AMD, Intel Arc, Apple Silicon and CPU.
SD.Next is an open-source server application and web interface for generating images and video with Stable Diffusion and many newer models. It suits people with non-Nvidia hardware or a Mac who still want a local, extensible interface.
What it is
You clone the repository, launch it with webui.sh, webui.bat or webui.ps1, and its built-in installer manages dependencies and updates. The README lists support for dozens of models and a Balanced Offload feature for running larger models on limited hardware.
Who it's for
It suits tinkerers who want one interface across several GPU vendors, including AMD ROCm or ZLUDA and Intel Arc. It does not suit anyone after a one-click consumer app or a hosted service with support.
Key features
Many GPU backends
CUDA, ROCm, ZLUDA, Intel IPEX, OpenVINO, DirectML and Apple MPS.
Dozens of models
One interface covers a broad, frequently updated model list.
Balanced Offload
Moves model parts to run larger models on limited VRAM.
Built-in installer
Automatic updates and dependency management on launch.
Image and video
Generates both still images and video clips.
Docker recipes
Container recipes for CUDA, ROCm, Intel IPEX and OpenVINO.
Up and running in 3 steps
Our honest verdict
What's great
- Runs on Nvidia, AMD, Intel Arc, Apple Silicon and CPU-only machines, which most local image tools do not cover.
- Apache-2.0 licence for the code, so you can use it in commercial work.
- Installer handles dependencies and updates itself after you clone the repo.
- Docker recipes exist for four GPU stacks if you prefer containers.
Where it falls short
- Needs a local GPU for reasonable speed; the README gives no minimum VRAM, so check each model's own requirements.
- Model weights are downloaded separately and each has its own licence you must read.
- The README points to external docs for most setup details, so first install takes reading.
- Not a hosted service: you handle updates, drivers and disk space yourself.
Pricing & license
Free under Apache-2.0. The only cost is the hardware and electricity you run it on; the project shows no paid hosted plan.
Open remote AI-training roles
870 open remote AI-training roles, updated November 14, 2026; most pay $20–$170 an hour. Zekai lists them free; the application goes to the employer’s platform.
Alternatives
These reviewed AI tools do the same job instantly — from our scored directory.
Use it, or have us set it up
FAQ
Does SD.Next work on AMD or Intel GPUs?
Yes. The README lists AMD GPUs through ROCm on Linux and Windows or ZLUDA on Windows, and Intel Arc through OneAPI with IPEX on Windows and Linux.
Can I use SD.Next commercially?
The code is Apache-2.0. Models you load carry their own licences, so check each one before commercial use.
Can it run without a GPU?
The README mentions CPU-only execution, but expect slow generation; a GPU is the normal way to use it.
