Draw Things: AI Generation
Draw Things: AI Generation is a user friendly AI-assisted image generation tool for iOS and Mac OS X 12.4 or later. Users can easily create...
Last verified:
Visit Draw Things: AI Generation
What is Draw Things: AI Generation?
Draw Things: AI Generation is a creative AI‑assisted image‑generation app that lets you run advanced models like Stable Diffusion and other large‑scale image and video generators directly on your own device, including iPhone, iPad, and Mac. It turns your text prompts into images—or even video—while keeping all processing local, so your data never leaves your device. You can describe a scene, choose a model and style, and then watch the app generate art in seconds or minutes, depending on complexity and your hardware.
The tool is built around local AI execution, which means it supports offline use and strong privacy by default. You can install and manage different backend models such as HiDream, Wan2.1, Hunyuan Video, Stable Diffusion 3 large, FLUX.1, and others, and then combine them with features like LoRAs, textual‑inversion embeddings, ControlNet, and TeaCache‑accelerated generation. The app also offers on‑device LoRA training so you can fine‑tune models yourself without uploading anything to a cloud server.
Draw Things is aimed at hobbyists, artists, and tech‑savvy creators who want deep control over AI image generation without installing command‑line tools or relying on third‑party servers. It suits people who value privacy, want to experiment with multiple models and techniques, and plan to run AI on Apple Silicon Macs or A12‑chip‑based iPhones and iPads. Whether you are making concept art, mood boards, or custom visuals for personal projects, the app bundles many advanced options into a single native interface.
Draw Things: AI Generation pricing
Pricing model: Free
Draw Things: AI Generation is free to download and run on iPhone, iPad, Mac, and Apple Vision devices, with local image and video generation available at no cost. The app description indicates that the base product does not require any subscription for core model support and generation features. There is at least one in‑app purchase tier called Draw Things+ priced at 8.99 USD, which offers additional functions or premium enhancements beyond the free tier, though the exact inclusions are not detailed in the App Store listing.
Draw Things: AI Generation pros
- Runs advanced image and video models locally on device
- Supports popular models like Stable Diffusion, HiDream, Wan2.1, Hunyuan Video, SD3, and FLUX.1
- No data collection; generation happens entirely on your device
- Provides on‑device LoRA training so you can fine‑tune models privately
- Offers a wide range of settings and well‑explained controls for advanced users
- Includes LoRA, textual‑inversion, and additive‑model support
- Supports ControlNet and video‑generation extensions
- Allows transparent‑image LoRAs for FLUX.1 workflows
- Integrates TeaCache acceleration to speed up repeated generations
- Comes with a revamped onboarding flow that guides new users
- Available for iPhone, iPad, Mac, and Apple Vision devices
- Optimized for Apple Silicon and metal‑based GPU usage
- No mandatory subscriptions or sign‑ups for basic image generation
- Free to download and use for core features
- Includes developer‑maintained updates and frequent feature additions
Draw Things: AI Generation cons
- Requires relatively recent hardware (A12 Bionic or Apple Silicon and iOS/iPadOS 15.4+ or macOS 12.4+)
- Can be complex and intimidating for beginners unfamiliar with Stable Diffusion
- User interface has known limitations on high‑resolution screens and in history browsing
- Lacks some UI conveniences such as smooth scrolling and bulk history navigation
- Limited options for managing long‑running or multi‑batch generations
- No built‑in cloud‑sync or cross‑device project sharing
- History and image grids can appear pixelated compared to main previews
- Some advanced features require manual configuration and experimentation
Frequently asked questions about Draw Things: AI Generation
Can I use Draw Things completely offline?
Yes, Draw Things can run entirely offline because it executes image and video models locally on your device. Once you download the required model files, the app does not need to connect to any remote server to generate images or videos, which helps protect your privacy and lets you work without an internet connection.
What types of models does Draw Things support?
Draw Things supports several major AI image and video models, including Stable Diffusion variants, HiDream, Wan2.1, Hunyuan Video, SD3 Large, FLUX.1, and others. It also allows on‑device LoRA training and supports ControlNet, textual‑inversion embeddings, and additive models so you can combine multiple styles and controls within the same workflow.
Does Draw Things collect my data or send images to the cloud?
According to the developer, Draw Things does not collect any data from the app, and all AI computations are performed locally on your device. This means your prompts, images, and any model‑training data stay on your iPhone, iPad, or Mac and are not uploaded or shared with external servers.
Is Draw Things suitable for beginners or only for experts?
Draw Things is designed with advanced users in mind, particularly those already familiar with Stable Diffusion and local AI workflows. It offers a lot of low‑level settings and customization, which can be overwhelming for beginners; however, the revamped onboarding flow and detailed inline explanations help newer users learn the basics and gradually explore more complex features.
What Apple devices can run Draw Things?
Draw Things supports iPhone devices with iOS 15.4 or later and an A12 Bionic chip or newer, iPad devices with iPadOS 15.4 or later and an A12 Bionic chip or newer, Macs with macOS 12.4 or later, and Apple Vision devices with visionOS 1.0 or later and an A12 Bionic chip or later. Performance is best on Apple Silicon‑based Macs and recent iPhones or iPads.
What is Draw Things+ and what does it include?
Draw Things+ is an optional in‑app purchase tier priced at 8.99 USD that unlocks additional features or enhancements beyond the free version. The exact benefit set is not fully itemized in the App Store listing, but such tiers in similar tools typically provide expanded model access, priority or faster generation modes, extra UI tools, or expanded export options.
Can I train my own LoRAs with Draw Things?
Yes, Draw Things supports on‑device LoRA training so you can fine‑tune models directly on your iPhone, iPad, or Mac. This lets you adapt a base model to your preferred style or subjects without uploading your data to any cloud service, preserving privacy while still giving you control over outputs.
How does Draw Things handle image resolution and quality?
The app lets you choose from several resolutions and model‑specific settings, and it is tuned to work best with 512×512‑style inputs. You can select different samplers and step counts to trade off speed versus quality, and TeaCache‑accelerated options can reuse parts of previous calculations to speed up re‑generations at similar resolutions.
Is there a way to stop an entire batch of generations at once?
Yes, the developer notes that you can stop all ongoing generations by right‑clicking the Skip button, which brings up a stop‑all option. This helps cancel long sequences or mistakes without having to manually skip each individual image in a batch.
Why are the history images sometimes pixelated on large displays?
On high‑resolution screens such as 4K monitors, the history sidebar shows small thumbnails that are scaled up, which can make them appear pixelated compared to the higher‑resolution previews in the main view. The developer has acknowledged this limitation and plans UI improvements to make history browsing smoother and more visually consistent.