What Is Forge UI Neo?
The original Stable Diffusion Forge Web UI was one of the most popular AUTOMATIC1111-based frontends — it added memory optimizations, faster generation, and a cleaner workflow on top of the classic SD interface. After its creator stepped away from active development, the project went quiet for a while.
Forge UI Neo is a community-driven fork that picks up where the original left off — and significantly expands it. This isn't just a maintenance update; it's a genuine new generation of the tool with support for the latest models and generation types.
Installation — Windows (Manual)
Prerequisites
- Python 3.11 — download from python.org (link in Resources below). Other versions may cause dependency issues.
- Git — download from git-scm.com if not already installed.
- Clone the Neo Branch — create or navigate to a folder where you want to host the project. Open a command terminal in that folder and run the clone command from the Forge Neo GitHub repository. Use the exact command from the repo — it specifies the Neo branch.
- Run webui-user.bat — open the newly created folder and double-click
webui-user.bat. On the first run, this kicks off the full installation — downloading and configuring all dependencies. This takes 15–20 minutes depending on your connection. On subsequent launches it starts Forge UI Neo directly. - Access the Web UI — once installation completes, Forge Neo automatically opens in your browser. If it doesn't, navigate to
http://127.0.0.1:7860.
Running Flux Dev Nunchaku — Example Workflow
Flux Dev Nunchaku is one of the headline additions in Forge Neo — it's currently the fastest way to run Flux Dev while maintaining excellent quality, thanks to its unique quantization approach.
Download and Place the Model Files
| File | Destination Folder |
|---|---|
Flux Dev Nunchaku model (.safetensors) | models/Stable-diffusion/ |
| CLIP L text encoder | models/text_encoder/ |
| T5 FP8 text encoder | models/text_encoder/ |
Flux VAE (ae.safetensors) | models/VAE/ |
Configure the UI for Flux
- Select the Flux UI Preset — in the top menu, find the UI Presets dropdown and select Flux. This reconfigures the interface for the Flux workflow.
- Select Your Checkpoint — choose the Flux Dev Nunchaku model from the main checkpoint dropdown.
- Select VAE + Text Encoders Together — in Forge Neo, VAE and text encoder models share a single selector called VAE / Text Encoders. Select all three at once: Clip L, T5 FP8, and the Flux VAE.
- Add LoRAs (Optional) — place LoRA files in
models/Lora/. In the generation settings, open the Lora submenu, hit refresh, then click your LoRA to automatically add its tag to your prompt. - Write Your Prompt and Generate — enter your prompt, set your resolution and sampling method, and click Generate.
📦 Want to skip the setup?
The Local Lab offers pre-configured AI installer packages so you can get running in minutes, not hours.
Get the Installer →
