No dedicated GPU
Prompt building, project organization, and standard non-AI editing are expected to remain usable. Local AI generation is not currently certified as a CPU-only launch target.
Compatibility & system requirements
Power when you want it. Clear guidance when you need it.
PeelCraft uses a verified development registry for supported engine profiles. The planned public PeelCraft installer ships with zero third-party AI model weights bundled; Model Manager is designed to direct users to official upstream sources identified by PeelCraft and verify the supported profile separately.
Development guidance is conservative and can change before public release. The verified system and provisional tiers below are not blanket performance guarantees.
Verified
Verified development configuration
The complete current nine-profile development registry has been exercised on this system. This is the verified quality target, not the final minimum specification.
- Windows 11 64-bit
- NVIDIA GeForce RTX 3090 with 24 GB VRAM
- 31.9 GB system RAM
- CUDA 12.6
- Nine installed development model profiles
- Automatic memory retry and model reuse enabled
GPU and VRAM guidance
Local generation requirements depend on the selected engine. Lower-memory operation may use offloading or other memory-saving techniques, but those options can be much slower and are not yet certified across the full PeelCraft workflow.
8 GB VRAM
Provisional lower target for SDXL Base, SDXL Turbo, and selected SDXL-family profiles. Final PeelCraft support depends on clean-machine testing, resolution, memory-saving mode, and model.
13 GB VRAM
FLUX.2 Klein 4B is designed for an approximately 13 GB consumer-GPU footprint upstream. The complete PeelCraft workflow at this tier still requires controlled certification.
16 GB VRAM
Current registry minimum for FLUX.1 Schnell and Z-Image Print Quality. This is the planned lower target for broader local generation, not yet a universal guarantee.
24 GB VRAM
Verified and recommended for the current full 1024-pixel quality development profile across all nine registered engines.
Operating system and processor guidance
The current application and local AI runtime are verified on one Windows and NVIDIA CUDA configuration. Other hardware remains pending until it is tested, not merely assumed.
Windows 11 64-bit
Current verified desktop target.
Windows 10
Not yet certified for the public release.
NVIDIA CUDA
Current verified acceleration path for local generation.
AMD and Intel graphics
Not yet certified for PeelCraft local generation. The website will not promise support before controlled testing.
CPU requirement
Exact processor minimum is not yet published. Standard editing and prompt workflows are expected to be lighter than local AI generation.
System memory and storage planning
Local AI models are large. Storage should be planned separately from the PeelCraft application itself.
32 GB system RAM
Verified and recommended for the current full development configuration.
16 GB system RAM
Provisional target for Smart Generator, standard editing, and selected lighter local-generation workflows. Not certified for all registered engines.
Model storage
Primary model packages in the current registry range from roughly 6.9 GB to 24.4 GB each. The shown development machine currently uses about 187.3 GB for nine installed models.
Free working space
Additional room is required for downloads, temporary files, generated artwork, previews, backups, and updates. Exact public minimum remains pending.
What PeelCraft is not promising yet
These items stay clearly unverified until they pass controlled testing.
CPU-only local generation
Not a certified launch target.
Below 8 GB VRAM
Not a certified local-generation target.
AMD or Intel local-generation acceleration
Not yet certified.
Windows 10
Not yet certified.
All laptop thermal and power configurations
Performance can vary and requires real-device testing.
All driver versions
Supported driver guidance will be published with the public compatibility matrix.
Current development engine registry
These are compatibility profiles, not PeelCraft-owned engines. The public PeelCraft installer is planned to bundle zero third-party model weights.
FLUX.1 Schnell
Development profileblack-forest-labs/FLUX.1-schnell- PeelCraft profile
- 1024 × 1024 · 4 steps · guidance 0
- Hardware guidance
- 16 GB registry minimum · 24 GB recommended
- Upstream license label
- Apache 2.0
FLUX.2 Klein 4B
Development profileblack-forest-labs/FLUX.2-klein-4B- PeelCraft profile
- 1024 × 1024 · 4 steps · guidance 1
- Hardware guidance
- 13 GB upstream target · 24 GB recommended
- Upstream license label
- Apache 2.0
Z-Image · Print Quality
Development profileTongyi-MAI/Z-Image- PeelCraft profile
- 1024 × 1024 · 36 steps · guidance 4
- Hardware guidance
- 16 GB registry minimum · 24 GB recommended
- Upstream license label
- Apache 2.0
SDXL Base 1.0
Development profilestabilityai/stable-diffusion-xl-base-1.0- PeelCraft profile
- 1024 × 1024 · 30 steps · guidance 5
- Hardware guidance
- 8 GB registry minimum · 12 GB recommended
- Upstream license label
- CreativeML Open RAIL++-M
SDXL Turbo
Development profilestabilityai/sdxl-turbo- PeelCraft profile
- 512 × 512 · 4 steps · guidance 0
- Hardware guidance
- SDXL lower-tier testing pending
- Upstream license label
- Stability AI Community License (repository label: sai-nc-community); registration, revenue, attribution, AUP, and commercial terms require review
DreamShaper XL
Development profileLykon/dreamshaper-xl-1-0- PeelCraft profile
- 1024 × 1024 · 28 steps · guidance 6
- Hardware guidance
- 8–12 GB provisional SDXL-family target
- Upstream license label
- CreativeML Open RAIL++-M
Juggernaut XL v9
Development profileRunDiffusion/Juggernaut-XL-v9- PeelCraft profile
- 1024 × 1024 · 30 steps · guidance 5.5
- Hardware guidance
- 8 GB upstream model-card target; PeelCraft compatibility validation pending
- Upstream license label
- CreativeML Open RAIL-M; paid API restriction applies
RealVisXL V5
Development profileSG161222/RealVisXL_V5.0- PeelCraft profile
- 1024 × 1024 · 30 steps · guidance 5
- Hardware guidance
- 8–12 GB provisional SDXL-family target
- Upstream license label
- CreativeML Open RAIL++-M
Animagine XL 4.0
Development profilecagliostrolab/animagine-xl-4.0- PeelCraft profile
- 1024 × 1024 · 28 steps · guidance 5
- Hardware guidance
- 8–12 GB provisional SDXL-family target
- Upstream license label
- CreativeML Open RAIL++-M
Important: PeelCraft is independently developed. Compatibility references do not create ownership, affiliation, sponsorship, partnership, certification by a model publisher, or endorsement. PeelCraft’s public installer is planned to contain zero third-party AI model weights; upstream licenses and current repository terms control any model a user chooses to obtain separately. Read the complete third-party technologies notice.