Draw Things: AI Generation App Reviews

I’m surprised that these developers made this app for free. No trials, no forced subscriptions, no annoying ads.

No rules, go crazy brother lol

Thanks, very good integration tools and it works very well. A minor bug is that it recognizes my M3 ultra Mac Studio as iPad, any choice to change this setting?

one of the best supported (if not the best) PRIVATE,LOCAL, image generation software you'll find out there. the software is frear, the developer is great, the community is AMAZING, supportive and engaging. i know this review needs to be improved but at least you know ai didn't write it lol. this is a great project worth checking it out

Crashes on certain things, like importing models and image to text. Other than that though, I’d say it’s pretty good and fun to use

The best way to run Local Image generation on phone and Mac. Your own models and terms. Unbeatable in easy with regular Updates and in depth settings.

This app is unbelievably feature-rich. I only have a small problem: when I download only an upscale model (real-esrgan4x) and then use the upscale preset, it causes the app to crash.

Drawthings is a great wrapper around most modern image/video generation models. This is an incredibly complex and nuanced landscape that drawthings does a pretty good job at rounding the edges of. Is it perfect? NO! But… if it were perfect, I wouldn’t have anything to whine about ;)

I may be biased but Draw Things is the complete package and is a must for anyone serious about creating with Stable Diffusion

Fantastic stable diffusion gui app

You can tell that this app began as an iOS app. A lot of the gripes I have for the app is how you can't queue work. For instance, downloading a model makes it impossible to do anything else, even downloading additional models. Also, there is no menu item for Settings, and the settings doesn't allow managing the models, Loras, etc. in one place. Perhaps having an activity queue would help the first issue, and having an asset manager under settings (and having a menu item to bring up the settings

选了个App内推荐的快速模型试了一下，出图速度确实很快

This is the best app I've used for image gen, both among web-based apps and native apps!

I may be biased but Draw Things is the complete package and is a must for anyone serious about creating with Stable Diffusion

I cannot believe this is a free app. This is by far the best UI for Stable Diffusion, and the performance is extremely good for how intensive Stable Diffusion is. The support for everything from loras ot control nets to... everything is so well designed. I love this app so much.

非常用心的 app! Mac的文生图 number one!操作简单,不需要非常繁琐的操作,开局下载好模型就可以使用!非常棒!大大降低了入门门槛!良心 app!强烈推荐!

Love this discovered it after Automatic1111 stop functioning on M1 macs. Big learing curve but really good results.

Overall a great app. Problem is now, it breaks the Files app. Cannot open projects with it Draw Things. Go to Files app and it constantly freezes and doesn’t show correct contents. Need to delete the app and all my projects in order to get a working Files app. Please fix this issue. This seems to be an iPad only issue that just began today for me.

Performance is much better than A1111 on macos m cpus... Would love to see InstantId and Pose for SDXL built into system.

Thank you, dev for doing whatever you did. The app is back to its original amazing working state.

The app crashes when I try to run image to image generation.

When using various models and Lora, the capacity of the Models folder becomes a burden. Please add an option to specify the Models folder.

Sdxl crashes on iPhone 15 pro

Draw Things is a fantastic and powerful frontend for running Stable Diffusion image generation and it’s COMPLETELY FREE. No in-app purchased, no ads, just a solid interface for running Stable Diffusion locally on Apple silicon. Runs excellently on my iPhone 13 Pro Max, runs even better on my iPad Pro, and I understand it runs extremely well on Macs as well. Don’t mess around with other AI image generators, just grab this one. The learning curve is probably a bit steeper than other apps, but the

This does a lot of stuff better than automatic 1111. The only thing that keeps this from being god-tier is that I don't think you can use automation scripts for things like animation or batches in this software.

Read above. I’m too lazy to write a description. All I could say is FIX IT!

This is the best application online for Mac when generation images with SDXL, bar none. Originally I had given up on using SDXL for regular image generation, but now my generation times on a mac M1 Pro are around a minute to a minute and a half. The application is not full of so many features like others interfaces, which has made things a little less hectic, but I still wish there was a little more documentation online for using it or youtube videos. Almost everything I found was through the Di

First downloaded this app and it was (if I recall correctly) basically just a Control Net copy, but opened it recently and wow! It’s truly rivaling the Automatic1111 desktop program but on my phone or iPad! Took a minute to figure everything out on my phone but on my iPad everything is in front of you to tweak any options. Really great work! Was completely surprised to see this supports Lora, TI, and checkpoints! Can’t wait to see how this progresses!

I've tried servral app such as DifussionBee and Diffusers. This app is only one that can work on my h-a-c-i-n-t-o-s-h. Great! awesome !

Okay app, but when you try to generate some images, it crashes, but mostly it works

looking forward to have more language, like chinese. And we can change words size

I hope to add content that supports Chinese.

I agree with Dragon knerd: I can’t seem to download any LoRA models from other websites like Civitai. Very frustrating. As for the existing LoRA already on the app, I have no idea what they produce unless I down load them. Some sort of quick tutorial would be nice for we beginners. Seed? What the heck is that? But I’m eager to learn and encouraged by all the positive comments!

Thanks to the developer for offering some Intel support. Seems to work. One problem: The GUI and text is TINY on a 27" 5K iMac.

I love this app, ever since I started venturing off into stable diffusion I’ve been looking for way to use it on my M1 iPad Pro, and this is definitely it! It has so many features and parameters just like on my desktop! One thing I wish it had though is a way to select rather to fill or not fill mask, and also mask blurring!! Other than those 2 things I absolutely LOVE this app, thank you so much for providing this awesome app to us!

Thanks for creating a true free offline AI generation. However, I’ve seen Civitai has loads of Lora models. Therefore, It would be nice for you to add Lora to the app to see it for myself. In addition to that, can you add dreambooth where I can create my own model. I would appreciate adding both of these to the app. Thank You

Messed around with Google Colabs, this tops for easibiltiy and customizing. Great job

it has many options it does list available models to download the only downside it in the history it does not list what was used to make the image - prompt, seed value etc. why did I even mess with automatic1111 ;)

By far the most featureful and flexible Stable Diffusion toolkit available for iOS, iPad, and Mac. Incredible achievement in development, getting things to run well on limited RAM, and the pace of development and updates is super fast. Can't recommend highly enough.

This app simplifies Stable Diffusion and makes it easy to dig into image generation using local resources (i.e., your own computer rather than a server in the cloud). I can't believe it's free. Good stuff, recommended.

My gaming laptop barely runs SD and yet Liu Liu has done the impossible by bringing AI image generation to our fingertips. Great work with the implementation and the frequent updates. One question though, some of my tokens turn red, while I am aware of the 75 token limit, does Draw Things restart the token weight after 75 tokens or are they excluded? The red color gives me the impression that they are not in use.

Hello, I don’t know if this is a bug or an NSFW filter but, when I put certain text such as "A masked entity, solo, no face, hidden face, samurai sword, katana, kitsune mask, looking down, side view, on a cliff scenario, detailed mask, perfect held sword, perfect mask." After generating that image, nothing happened, it doesn’t give me any image for some reason. I don’t know what caused it, do can you explain why I can’t generate certain image like that? Thank you

I’ve used other UIs for making reference images. I loaded the same model I frequently use in Invoke and copied a prompt I had used with that to test. The image just turned into a gray square. The style I need is specific to that SDXL model. I love the speed, but not the results.

I dont know how old the models are for this app, but after trying this out extensively, this app leaves a lot to be desired. It constantly produces extra limbs, fingers, body parts as well as twisted legs and multiple heads even when adding all of this to negative prompts. I dont mind using the advanced tools to tweak images, but you can on ly do so much with poor generations.

The most recent version of the app has a issue where the app immediately stops and exits after starting a generated process. Meaning that the app is unusable for the moment. Could you please fix this?