ArtCraft Download: Official Release Channels & Setup

ArtCraft is free, open-source software (MIT / Apache-2.0). We maintain this independent installation index to direct creators to safe, verified upstream GitHub binaries and official checksum records.

Official Channel Advisory:

Download ArtCraft exclusively through the official GitHub Releases repository (storytold/artcraft) or getartcraft.com. Beware of unofficial repackaged mirror portals. Always inspect the publisher signatures and compare checksums directly against official release tables.

OFFICIAL MSI & EXE

Windows 10 / 11 (64-bit)

ArtCraft_0.41.0_x64-setup.exe / .msi

MINIMUM HARDWARE:NVIDIA RTX 2060 / 3060 (6GB+ VRAM)

RECOMMENDED:RTX 3070 / 4070 (8GB–12GB+ VRAM)

ENVIRONMENT:NVIDIA Studio Driver 550.54+ with CUDA support

INTEGRITY VERIFICATION:

Verify SHA-256 via official GitHub Releases assets table

~42 MB – ~47 MB (Installer)Official Release

UNIVERSAL DMG (M1–M4)

macOS Sonoma / Sequoia (Apple Silicon)

ArtCraft_0.41.0_universal.dmg

MINIMUM HARDWARE:Apple M1 with 16GB Unified Memory

RECOMMENDED:Apple M2/M3/M4 Pro or Max (24GB+ Unified Memory)

ENVIRONMENT:Apple Metal 3 / macOS 14.2+ native execution

INTEGRITY VERIFICATION:

Verify SHA-256 via official GitHub Releases assets table

~95 MB (Universal DMG)Official Release

BUILD FROM SOURCE / DOCKER

Linux (Ubuntu 22.04+ / Debian / Arch)

Compile via storytold/artcraft source tree

MINIMUM HARDWARE:NVIDIA RTX or AMD Radeon RX 6800+ (ROCm 6.1)

RECOMMENDED:NVIDIA RTX 3090 / 4090 with 24GB VRAM

ENVIRONMENT:Docker / PyTorch 2.3+ with CUDA or ROCm libraries

INTEGRITY VERIFICATION:

Follow official upstream build instructions on GitHub

Source Tree / Container ImageOfficial Release

SETUP & TROUBLESHOOTING

Common First-Launch Errors & Solutions

CUDA Initialization Failed / GPU Not Detected

Ensure you have installed the NVIDIA Game Ready or Studio Driver version 550.x or later. If running in a virtual machine or WSL2, verify that NVIDIA Container Toolkit is installed and CUDA pass-through is active.

Apple Silicon Out-of-Memory (MPS Allocation)

On 8GB/16GB unified memory Macs, running SDXL alongside background apps may exceed unified memory limits. Enable the "Low VRAM Mode" toggle in ArtCraft Settings > Performance to stream model weights sequentially.