ComfyUI
Node-based interface for local image and video generation
Draw Things does something genuinely unusual: it runs full Stable Diffusion and Flux models entirely on an iPhone, iPad or Mac, offline, for free, with your prompts never leaving the device. The boundary is in the name of its maker ecosystem — it is Apple-only by design, built on Metal and Apple silicon, with no Windows, Android or browser edition. It also asks you to download models and manage local storage, and your generation speed is whatever your device can give you. If you are outside the Apple world, or you want more control or more horsepower than a phone GPU, these are the alternatives worth your time.
On Windows or Linux, the closest equivalents are ComfyUI (maximum control, node-based workflows) and Stable Diffusion WebUI Forge (the familiar WebUI, faster and lighter on VRAM). For the least setup, Fooocus strips local generation down to a prompt box. And if "local" itself is the problem, Midjourney removes the hardware question entirely for a subscription. Inside the Apple ecosystem, Draw Things is hard to argue with — it is free, private and genuinely capable. Outside it, pick by temperament: Forge for the classic WebUI done faster, ComfyUI for pipelines you can inspect and repeat, Fooocus or SwarmUI for a gentler start, and Midjourney when you would rather rent the hardware than own the workflow. All four local picks run the same open model families, so the switching cost is mostly interface, not assets.
| Tool | Pricing | Free to use | API | Open source | Compliance |
|---|---|---|---|---|---|
| ComfyUI | Free / Pay-per-use | Yes | — | — | — |
| Stable Diffusion WebUI Forge | Free | Yes | Yes | — | SOC 2 |
| Fooocus | Free | Yes | — | — | — |
| SwarmUI | Free | Yes | Yes | Yes | SOC 2 |
| Midjourney | Paid | No | Yes | — | — |
A dash means the vendor does not state this publicly.
Node-based interface for local image and video generation
The power user's answer, and the one Draw Things itself points Windows users to. ComfyUI replaces the app interface with a node graph: every model, sampler and parameter is a box you can inspect, rewire and save, which makes workflows reproducible in a way a mobile app cannot match. It runs locally and free on Windows, Mac and Linux. There is also an official cloud — a free tier with a handful of GPU runs, then Standard at $16 a month billed yearly, Creator at $28 and Pro at $80 — for days when your own hardware is busy. The cost is a learning curve measured in evenings, not minutes.
Stable Diffusion WebUI Forge is a platform built on top of Stable Diffusion WebUI that optimizes resource management, speeds up inference, and makes experimental features easier to study. It targets users who run Stable Diffusion locally and want better performance and memory handling without abando
The familiar-interface route. Stable Diffusion WebUI Forge keeps the AUTOMATIC1111 WebUI layout that most tutorials teach, then rewrites memory handling underneath: independent benchmarks put it around 30% faster than the original on SDXL, and its native support for quantized Flux models (NF4 and GGUF) brings Flux down to roughly 6–8 GB of VRAM. It is free, open source, and installs from a one-click package. Choose it over Draw Things when you have an NVIDIA card and want your LoRAs and extensions to behave the way the guides describe.
Midjourney-style local image generation with almost no settings
Local generation with the settings removed. Fooocus is built on SDXL and aims squarely at people who found the WebUI intimidating: type a prompt, pick a style, and it manages the parameters Midjourney-style behind the scenes. It is free, offline and open source, distributed from GitHub. What you give up is depth — no node graphs, a narrower model story than Forge or ComfyUI — and the project describes itself as non-commercial software, so check the licence before using outputs in paid work.
SwarmUI is a free, open-source web interface for running AI image, video, and audio generation models locally, aimed at both beginners and advanced users who want accessible power tools and extensibility.
The middle path. SwarmUI is a free, open-source local interface designed to be approachable on day one and deep later: a friendly generate tab up front, with fuller workflow control available as you grow into it, and support for image, video and audio models in the same install. Against Draw Things, the trade is clear — you lose the App Store install and the phone in your pocket, and you gain a desktop tool that runs on the hardware you already own, Windows and Linux included.
Midjourney is an AI image and video generation platform built by a self-funded, distributed lab of about 60 people. It targets artists, designers, and creative teams who want stylized visuals and short video clips from text prompts, with tiered subscriptions that scale from limited monthly jobs to u
The cloud exit. If the real obstacle is not the operating system but the hardware — no GPU worth the name, no patience for model downloads — Midjourney moves generation to its own servers for a subscription: Basic at $8 a month on annual billing, Standard $24, Pro $48, Mega $96. Output quality and house style are the draw; what you give up is everything Draw Things stands for: offline use, on-device privacy, local LoRA training and a free tier. Your prompts and images live on someone else's computer, at someone else's prices.
No. Draw Things is built exclusively for iPhone, iPad and Mac around Apple silicon and Metal; there is no Windows, Android or web version. The closest Windows equivalents are ComfyUI and Stable Diffusion WebUI Forge, which run the same open models on NVIDIA and AMD cards.
ComfyUI, Forge, Fooocus and SwarmUI are all free to run locally; ComfyUI's optional cloud is the paid part. Midjourney is subscription-only, starting at $8 a month billed annually. Draw Things itself is free for local generation, with an optional Draw Things+ plan at $8.99 a month for higher cloud limits.
Fooocus, by design — it hides nearly all parameters. SwarmUI is the next step up, with a friendly default interface and more depth when you want it. Draw Things remains the easiest install of all, but only on Apple devices.
Largely, yes. The local tools here run the same open Stable Diffusion and Flux model families, and LoRAs trained or downloaded for one generally work in the others, subject to each tool's format support. Workflows and settings, however, do not transfer — only the model files do.