DiffusionBee is back.
Meet Diffusion4Mac.
The ultimate Local Flux macOS app. Run state-of-the-art AI models like FLUX.1 and Flux.2 [klein] directly on your Apple Silicon Mac—completely private, offline, and sub-second. It's the ultimate DiffusionBee update for 2026.
Compatible with M1, M2, M3, M4, M5 Apple Silicon Macs. Requires macOS 12.3+
Designed for Next-Gen Local AI
Experience the raw performance of a native Flux Klein Apple Silicon GUI combined with flexible cloud integrations.
Local Power
Achieve sub-second generations directly on your machine. Fully optimized for Apple Silicon GPUs and Neural Engines using FLUX.1-schnell and FLUX.2 [klein].
Hybrid Flexibility
Seamlessly transition between offline workflows and cloud-based models. Use the Gemini Studio tab for premium fallback to Google's Imagen 3 and prompt engineering.
Onboard Intelligence
Unleash local prompt enhancements and image assessments with the built-in, local Gemma-powered AI Assist and Prompt Designer. Run Gemma locally for purely local prompting and image assessment completely offline.
Advanced Control
Fine-tune your generations with multi-LoRA stacking. Import custom weights safely with automatic trigger word extraction and required keyword validation.
An All-in-One Desktop AI Studio
Diffusion4Mac packs high-performance capabilities into a clean, unified sidebar interface.
Text to image
Generate high-quality images from text descriptions locally using FLUX models.
Image to image
Transform existing images using structural guidance and prompt descriptions.
Inpainting
Add, remove, or modify objects within a specific region of your images seamlessly.
Upscaler
Increase the resolution and add fine detail to low-resolution generations locally.
Prompt Designer
Design complex, artistic FLUX prompts using local vision models.
LoRAs
Manage, validate, and apply custom fine-tuning weights with trigger word checks.
Gemini Studio
Access cloud-based frontier AI generation utilizing Imagen 3 and Gemini 2.5 Flash.
Engineered to Run Flux Locally on Mac
Diffusion4Mac is compiled natively to harness Metal Performance Shaders (MPS) and Apple Silicon’s unified memory architecture.
- Fully Offline: Run Flux locally on Mac without internet requirements.
- Zero Telemetry: Complete privacy. Your generations never leave your hardware.
- Resource Optimized: Low RAM configurations enable FLUX on baseline Macs.
- Model Formats: Supports .safetensors checkpoints, custom VAEs, and custom LoRA files.
Technical Matrix
| Parameter | Requirement / Value |
|---|---|
| Minimum OS | macOS Monterey (12.3) or newer |
| Architecture | Apple Silicon (M1/M2/M3/M4/M5) Native |
| Base Models | FLUX.1-schnell, FLUX.2 [klein] |
| Cloud Fallback | Gemini Studio (Imagen 3 & Gemini 2.5 Flash API) |
| Licensing | GNU AGPL-3.0 (Derivative of DiffusionBee) |