Draw Things

Draw Things is an AI-assisted image generation app for iOS and macOS, leveraging Stable Diffusion models to create and edit images offline. Available on the App Store and official website, it supports text-to-image, image-to-image, inpainting, and video generation, all processed locally to ensure user privacy. The app is free for local use, with an optional Draw Things+ paid tier for cloud compute.

Key features include Text-to-Image, which generates images from prompts like “a futuristic city at dusk” in minutes, and Inpainting Mask, which allows users to edit specific image areas. Infinite Canvas enables endless image expansion, while Scribble-to-Image transforms rough sketches into detailed visuals. Video generation, supported by models like Hunyuan, produces short clips. The app supports community models and LoRAs, such as Waifu Diffusion v1.3 for anime or Stable Diffusion v1.5 for inpainting, and is optimized for Apple Silicon (M1/M2/M3).

Performance depends on device specs. On an M1 Mac, a 1024×1024 image with FLUX.1 dev takes about 3-4 minutes at 20 steps, slower than A1111 Stable Diffusion WEB UI on high-end PCs. Users praise the intuitive interface and offline privacy but note occasional iCloud syncing issues with LoRAs and slower rendering on older devices like iPhone 11. Compared to MidJourney, which requires cloud access, Draw Things prioritizes privacy and ease of use. DiffusionBee is a similar offline option but lacks video support.

The app’s community on Discord and Reddit provides robust support, with guides and model recommendations. Recent updates, like version 1.20250708.0, improve video frame consistency and add TeaCache for faster rendering. Draw Things suits artists and casual users seeking a private, accessible AI art tool.

For best results, use on M1/M2 devices, start with simple prompts, and explore community LoRAs for specialized styles. Check the wiki for setup guides and join Discord for troubleshooting.