Draw Things
Draw Things develops an AI image and video generation app that runs locally on Apple devices.
Founded
2023Draw Things develops an AI image and video generation app that runs locally on Apple devices.
Draw Things offers an offline AI creation app for iPhone, iPad, and Mac. The product generates images and video using diffusion models including SDXL, FLUX, and LTX-2.3, with support for text-to-image, image-to-video, style transfer, inpainting, outpainting, ControlNet, pose editing, and on-device LoRA training.
The app operates entirely on-device for privacy, requiring no internet connection for core generation tasks. Optional cloud processing and subscription tiers are available for users who need additional compute capacity beyond what their local hardware can provide.
Draw Things operates entirely on-device for privacy, requiring no internet connection for core image generation. The app runs natively on Apple Silicon for optimal performance, supports a wide range of open diffusion models, and offers advanced features like LoRA training and ControlNet that competitors typically restrict to desktop or cloud platforms.
Unlike cloud-based image generators such as Midjourney or DALL-E, Draw Things keeps all user data and generated content on the local device. This privacy-first approach, combined with free offline access and no token limits, distinguishes it from subscription-only competitors in the generative AI space.
Draw Things offers a free tier for local offline generation with no advertisements or token limits. Users can download and use the core image generation features at no cost. Optional cloud processing and a Draw Things+ subscription at $8.99 per month provide additional compute capacity and premium features.
Boost bundles are available for one-time purchases ranging from $0.99 to $19.99, offering flexible options for users who need occasional extra processing power without committing to a monthly subscription. This freemium model makes advanced AI image generation accessible to casual users while supporting power users with scalable compute options.