A polished local image generation studio powered by Automatic1111 and Stable Diffusion. Juggernaut Reborn for photorealism, DreamShaper 8 for artistic work. Your hardware, your models, zero cloud.
The App
A clean, distraction-free frontend for Automatic1111 (A1111) — the industry-standard local image generation server. Switch between two pre-configured models, build prompts with one-click style chips, and control every generation parameter without leaving the app. Generated images stay on your machine.
Automatic1111 needs Python and Git installed. If you already have them, skip straight to Step 2.
Python 3.10.6 (exact version recommended for A1111):
# Download Python 3.10.6 installer from:
https://www.python.org/ftp/python/3.10.6/python-3.10.6-amd64.exe
# ✅ During install, check "Add Python to PATH"
# ✅ Choose "Install Now"
# Install Git from:
https://git-scm.com/download/win
# Accept all defaultspython --version → Python 3.10.6git --version → git version 2.x.x# Install Homebrew if you don't have it
/bin/bash -c "$(curl -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh)"
# Install Python and Git
brew install [email protected] gitpython3 --version → Python 3.10.x · git --version → git version 2.x.x# Ubuntu / Debian
sudo apt update
sudo apt install python3.10 python3.10-venv git wget
# Fedora / RHEL
sudo dnf install python3.10 git wgetpython3.10 --version → Python 3.10.xDownload all three files into the same folder — the startup script expects to find the app file in the same directory.
C:\Atelier\ or ~/atelier/ and put all three files there. The startup script will clone Automatic1111 into a subfolder automatically on first run.Your folder should look like this:
Atelier/
├── atelier.html ← the UI app
├── START_A1111.bat ← Windows launcher
└── start_a1111.sh ← Mac/Linux launcher
# stable-diffusion-webui/ will be created automaticallyAtelier is pre-configured for two best-in-class Stable Diffusion models. Download both (or just one) from Civitai — free account may be required.
The go-to model for photorealistic portraits, landscapes, and product photography. Exceptional skin texture and lighting.
Excellent all-around model for fantasy, concept art, anime, and stylized work. Great at following complex artistic prompts.
Copy both .safetensors files to the models folder:
# Copy to (create the folder if it doesn't exist yet):
Atelier\stable-diffusion-webui\models\Stable-diffusion\
# Or let the script run first to create the folder structure,
# then copy the model files in.~/atelier/stable-diffusion-webui/models/Stable-diffusion/
# Example using terminal:
cp ~/Downloads/juggernaut_reborn.safetensors \
~/atelier/stable-diffusion-webui/models/Stable-diffusion/
cp ~/Downloads/DreamShaper_8_pruned.safetensors \
~/atelier/stable-diffusion-webui/models/Stable-diffusion/Run the included startup script. On first run it will automatically clone Automatic1111 from GitHub, install all Python dependencies, and start the server with the API and CORS enabled. This takes 5–15 minutes on first run.
# Double-click in Explorer, or from PowerShell:
cd C:\Atelier
.\START_A1111.bat# Make executable (only needed once)
chmod +x ~/atelier/start_a1111.sh
# Run
cd ~/atelier
bash start_a1111.shchmod +x ~/atelier/start_a1111.sh
cd ~/atelier
bash start_a1111.shRunning on local URL: http://127.0.0.1:7860With A1111 running, open the Atelier app in your browser. It connects automatically to 127.0.0.1:7860.
# Double-click atelier.html in Explorer
# Or from PowerShell:
start C:\Ateliertelier.htmlopen ~/atelier/atelier.html
# Or double-click in FinderNow generate your first image:
1. Select a model — Juggernaut Reborn or DreamShaper 8
2. Type a prompt (or click style chips to build one)
Example: RAW photo, portrait of a woman, cinematic lighting,
sharp focus, photorealistic, golden hour
3. Adjust settings if needed:
Steps: 20–30 for quality · CFG: 7 for most prompts
Resolution: 512×512 fast · 768×768 detailed
Hires. fix: ON for print-quality upscale
4. Click ⚡ Generate (or press Ctrl+Enter)
→ Images appear in the gallery below the form
→ Click any image to preview full-size
→ ↓ Download single · ↓ Save all → ZIPTroubleshooting
Most issues come from Python version mismatches, missing models, or the server not being started before opening the app.
A1111 isn't running or started on a different port.
# Make sure A1111 is running:
START_A1111.bat # Windows
bash start_a1111.sh # Mac/Linux
# Wait for: Running on local URL: http://127.0.0.1:7860
# The terminal window must stay openPython dependencies not installed. Let the script handle it:
# Delete the venv and re-run the script
rmdir /s stable-diffusion-webui env (Windows)
rm -rf stable-diffusion-webui/venv (Mac/Linux)
# Re-run the startup script — it will reinstallThe .safetensors file is missing or misnamed.
# Check the models folder:
stable-diffusion-webui/models/Stable-diffusion/
juggernaut_reborn.safetensors ← exact name
DreamShaper_8_pruned.safetensors ← exact name
# Filenames are case-sensitive on Mac/LinuxA1111 was started without the CORS flag. Use the included scripts — they pass --cors-allow-origins=* automatically.
# Stop A1111 (close the terminal window)
# Re-run using the provided script, not the default
# webui.bat / webui.sh directly
START_A1111.bat # ← always use thisRunning on CPU instead of GPU. Check CUDA installation.
# Check if GPU is detected:
cd stable-diffusion-webui
python -c "import torch; print(torch.cuda.is_available())"
# Should print: True
# If False: reinstall PyTorch with CUDA from pytorch.orgGeneration resolution too high for your GPU VRAM.
# Reduce resolution in Atelier:
4 GB VRAM → max 512×512
6 GB VRAM → up to 640×640
8 GB VRAM → up to 768×768
12 GB+ VRAM → 1024×1024 with Hires. fix
# Or add --medvram flag to the startup scriptHardware Guide