
Draw Things
@drawthingsapp
Followers
5K
Following
74
Media
464
Statuses
883
#stablediffusion in your pocket. 100% offline and free.
San Francisco, CA
Joined November 2022
🚨 Qwen Image Edit 2509 is FINALLY HERE! 🔄 Draw Things 1.20250930.0 was released in iOS / macOS AppStore a few hours ago. This version brings you: 1️⃣ Fix network issues connecting Cloud Compute on iOS 26; 2️⃣ Support Qwen Image Edit 2509, this is the first version of Qwen
6
14
163
4️⃣By the way, if you only use this detailed prompt without including picture 2 as a style reference, you still won’t get the ideal result. The outcome comes from the combined effect of both the prompt and the reference image. You can try it yourself and see the difference.
0
0
2
3️⃣So, you need to describe picture 2’s style in a more detailed and precise way. How to do that? I gave picture 2 to Qwen3-VL, and it did an excellent job of reverse-engineering the style for me. Then, I added the reversed style description to the second part of the prompt —
2
0
1
2️⃣If you only write the prompt as — “convert [picture 1] into the line-art style of [picture 2]” . you’ll most likely end up with an unsatisfying result — probably just a basic pencil sketch style, as shown below.
1
0
1
1️⃣In our previous post, we mentioned that QIE-2509 has built-in ControlNets, which allow it to use another image as a style reference. So, we placed Sam Altman on the Canvas and added a Nobel Prize winner’s line-drawing portrait to the Moodboard. In this setup, Sam Altman
1
0
2
🧪 It’s not perfect, but very close. 🔂 To replicate the style of the 2025 Nobel Prize winner without using any specifically trained LoRA. 🕹️ Just use Qwen Image Edit 2509 and prompts to achieve it. 👇🏻 Below are the detailed steps and reference prompts.
1
4
23
👍 I have to say, THIS IS AMAZING!! ❤️🩹 The ability to restore and recolor old photos — Qwen Image Edit 2509 performs exceptionally well. 🌇 In the four examples in Draw Things below, the latest Qwen Image Edit 2509 Lightning 4-Step LoRA was used — bringing old photos back to
2
19
138
6️⃣ From sketch to realistic photo — you can use another image as a style reference.
2
0
8
5️⃣ When the canvas is empty, place three images on the Moodboard. (Note: Please put the control image first.) Well, in this case, it works pretty well with three images.
2
0
9
4️⃣ When the control image is on the canvas, and there are multiple images on the Moodboard at the same time.
1
0
6
3️⃣ When the control image is on the canvas, and the subject image is in the Moodboard.
1
0
6
2️⃣ When the canvas is empty, and both the control image and the subject image are in the Moodboard.(Note: The weight of images in the Moodboard does not take effect.)
1
1
9
🕹️ Qwen Image Edit 2509 comes with built-in ControlNets. 👇🏻 Here are a few cases showing how to use them in Draw Things. 1️⃣ When the canvas is empty, and the control image is in the Moodboard:
9
33
268
🚨 A CASE to show you 2509’s Multi-Image Editing in Draw Things!! 🔢 Put images in the Moodboard, counted from left to right. If there is an image on the canvas: picture 1 = image on the canvas; picture 2 = first image in the moodboard; and so forth… 🧩 So, this way you can
1
5
42
2. Prompt2: 【Male】Generate an ID registration photo of this man, wearing a simple black suit over a white shirt, keeping his facial features unchanged, sitting upright, eyes looking directly at the camera, and a solid blue gradient background. Skin tone should appear natural
0
1
8
1. I’ll share two prompt templates for generating ID photos. You guys could try them out and customize the clothes, expressions, and backgrounds as needed. 【Male】Convert the picture into an American-style professional headshot in corporate photography style, while preserving
1
1
19
👩🏻💼 From a Casual Daily Photo to a Work ID Photo or Profile Picture. 🆚 Qwen Image Edit already does a decent job, but 2509 performs even better! 🧪 All the images were generated in Draw Things. 👨🏼💻 Qwen Image Edit 2509 Support is on the way, stay tuned! 👇🏻 Click on this post
3
8
88
3. After testing, Kontext with 8-step Turbo LoRA combined with the Apple Emoji LoRA can already produce good results. You could use the following parameters as a starting point (Copy and Paste): ——————
0
0
2
2. Prompt reference: {turn the scene and people into “apple emoji” style.} If you only want to turn the people into Apple Emoji while keeping the background pure white, you can write: {turn the people into ‘apple emoji’ style, the background should be pure white}
1
0
3
1. This is a Kontext LoRA specially trained for Apple Emoji, with the link provided below. It has shown good results when tested in Draw Things. If you’re a fan of Apple Emoji, this could be a great choice. Download it and import it into the Draw Things LoRA model list, and set
1
0
1