How do I install Forge Neo on Windows?

Two options. The one-click Windows installer downloads and configures the Forge Neo build, an isolated Python environment and the performance stack (Sage Attention, Flash Attention 2, Triton) for you. Or install manually: clone the Forge Classic Neo repository on the Neo branch, create a Python 3.11 environment, install the requirements, and optionally add Flash Attention and Triton.

What Python version does Forge Neo need?

Python 3.11 for manual installs. The one-click installer bundles its own isolated environment, so you do not need to manage Python yourself.

What are the system requirements for Forge Neo?

An NVIDIA GPU is recommended for the attention and Triton acceleration. Image generation runs comfortably on 6-8 GB of VRAM, and WAN 2.2 video targets low-VRAM cards via GGUF. Budget 30 GB or more of free disk space for the app plus models.

Yes, Forge Neo is open-source. The one-click installer is a convenience that pre-configures everything for you.

How to Install Forge Neo on Windows (2026)

Forge UI Neo (also called Forge Classic Neo) is the actively-maintained community fork of Stable Diffusion WebUI Forge — a fast, memory-efficient frontend for local AI image and video generation on Windows. There are two ways to install it: a one-click installer that configures everything for you, or a manual setup. Here is both, plus the requirements you need to hit either way.

Fast track: the quickest way to get Forge Neo running is the one-click Windows installer — it sets up Forge Neo with Sage Attention, Flash Attention 2 and Triton pre-configured, so you skip the manual dependency setup entirely.

Before you start — system requirements

GPU: an NVIDIA GPU is recommended for the attention/Triton acceleration.
VRAM: image generation runs comfortably on 6–8 GB; WAN 2.2 video also targets low-VRAM cards via GGUF. See the Local AI VRAM Requirements 2026 guide for model-by-model numbers.
Python: 3.11 for manual installs (the one-click installer bundles its own environment).
Disk: budget 30 GB+ of free space for the app plus models.

Option A — One-click installer (easiest)

The Forge Neo one-click installer downloads and configures everything for you:

The Forge Neo build itself
An isolated Python environment — nothing to install or match by hand
The performance stack: Sage Attention, Flash Attention 2 and Triton

You double-click and it runs — no command line, no dependency hunting. It is also available free with Local Lab Pro.

Option B — Manual install

If you prefer to set it up yourself:

Clone the Forge Classic Neo repository — on the Neo branch (see the warning below).
Create a Python 3.11 environment.
Install the requirements.
Optionally add Flash Attention + Triton for a speed boost.

This route works, but it is finicky on Windows — matching the right PyTorch, CUDA, Triton and attention versions is the part most people get stuck on. That version-matching is exactly what the one-click installer exists to skip.

💡 Neo branch vs Classic branch The repository has two branches. The Neo branch carries all the new features (WAN 2.2, Nunchaku, Flux Context) and is actively developed. The Classic/Main branch is an archive built on an older stable Forge release — if you are installing manually, make sure you clone the Neo branch.

After installing — what Forge Neo supports

Model	What it does	One-click setup
WAN 2.2 (T2V / I2V)	AI video — text-to-video and image-to-video, right inside Forge Neo	WAN 2.2 in Forge Neo installer
Z-Image Turbo	Fast, high-quality image generation	Z-Image Turbo in Forge Neo installer
Flux (incl. Nunchaku)	State-of-the-art image generation; Nunchaku for faster Flux	Built into the base Forge Neo installer
SDXL / SD 1.5	Classic Stable Diffusion checkpoints & LoRAs	Supported natively

For a full walkthrough of running Flux Dev Nunchaku — where each model file goes and how to configure the UI presets — see the complete Forge Neo guide.

FAQ

Is Forge Neo free? Yes — it is open-source. Our one-click installer is a convenience that pre-configures everything for you.

What Python version does it need? Python 3.11 for manual installs; the one-click installer bundles its own environment.

Does Forge Neo support WAN 2.2 video? Yes, both text-to-video and image-to-video.

How do I keep it up to date? See how to update Forge Neo.

Next steps: the complete Forge Neo guide covers extensions, model support and Forge Neo vs ComfyUI. Every Forge Neo installer is in the store or included with Local Lab Pro.