How to Run FLUX Locally on Windows, Mac, and Linux (Step-by-Step Guide)

How to Run FLUX Locally on Windows, Mac, and Linux

FLUX, a cutting-edge AI image generator, is making waves in the AI landscape with its ability to craft detailed and imaginative visuals from text prompts. While online versions offer convenience, running FLUX locally provides faster processing, enhanced privacy, and limitless usage. This guide will walk you through four simple methods to set up FLUX on your computer, whether you’re using Windows, Mac, or Linux.

Why Run FLUX Locally?

Running FLUX on your system ensures:

• Improved Privacy: Your prompts and generated images remain private.
• Faster Processing: Utilize your system’s GPU for quicker results.
• Unlimited Use: No restrictions on usage or features.

Methods to Run FLUX Locally

Method 1: Using ComfyUI

ComfyUI is an intuitive interface designed for running AI models, making it perfect for those who prefer graphical setups.

Steps to Set Up ComfyUI:

Download and Install ComfyUI

• Visit ComfyUI GitHub Releases.
• Download the latest version for your OS and extract it to a preferred location.

Download FLUX Models

• Access the FLUX HuggingFace Repository and download:
  • flux_schnell.safetensors (Main Model)
  • ae.safetensors (VAE File)
• From FLUX Text Encoders, download:
  • t5xxl_fp16.safetensors (for systems with 32GB+ RAM)
  • t5xxl_fp8_e4m3fn.safetensors (for systems with less than 32GB RAM)
  • clip_l.safetensors

Place Files in Relevant Folders

• Move the downloaded files into the corresponding ComfyUI folders:
  • models/checkpoints/
  • models/vae/
  • models/clip/

Run ComfyUI

• Open a terminal, navigate to the ComfyUI folder, and run:
  • Windows: python_embeded\python.exe -m ComfyUI
  • Mac/Linux: python3 main.py
• Open http://localhost:8188 in a browser.

Set Up FLUX Workflow

• Add and connect nodes in the interface (e.g., CLIP Text Encode, Flux Guidance).
• Input your text prompt, adjust settings, and generate images.

Method 2: Using Stable Diffusion WebUI

Stable Diffusion WebUI is a popular choice for users already familiar with its interface.

Steps to Set Up Stable Diffusion WebUI:

Install Stable Diffusion WebUI

• Clone the repository:

git clone https://github.com/AUTOMATIC1111/stable-diffusion-webui.git

• Navigate to the directory:

cd stable-diffusion-webui

• Run the appropriate script:
  • Windows: webui-user.bat
  • Mac/Linux: ./webui.sh

Download FLUX Models

• Follow the steps in Method 1 to download FLUX models and VAE files.

Configure Stable Diffusion WebUI

• Place FLUX files in the respective folders.
• Adjust settings in the WebUI to select FLUX models and VAE.

Generate Images

• Enter prompts, tweak settings, and create images seamlessly.

Method 3: Using Stability Matrix

Stability Matrix combines a sleek interface with advanced customization options for FLUX users.

Steps to Use Stability Matrix:

Install Stability Matrix

• Download the installer from the Stability Matrix GitHub page.
• Follow installation instructions based on your OS.

Set Up FLUX Models

• Use the in-built model manager to download and configure FLUX files.

Launch ComfyUI within Stability Matrix

• Configure nodes and workflows as outlined in Method 1.

Method 4: Using Python Script (Advanced Users)

This approach is ideal for those comfortable with coding and Python environments.

Steps to Set Up Python Script:

Prepare Python Environment

• Install Python 3.8+ and create a virtual environment:

python -m venv flux_env

• Activate the environment and install dependencies:

pip install transformers diffusers accelerate

Download FLUX Models

• Follow the steps in Method 1.

Create and Run Python Script

• Write a script to load models, process prompts, and generate images.
• Run the script to create visuals locally.

Conclusion

Running FLUX locally unlocks unparalleled opportunities for crafting high-quality AI-generated images. Whether you choose a user-friendly interface like ComfyUI or opt for advanced Python scripting, this guide equips you with all the tools you need. Experiment with prompts, fine-tune workflows, and embrace the creative potential of FLUX.