How to Play Around with ComfyUI? A Designer’s Fun Test
At the start of this year, I stumbled upon ComfyUI, and wow, this tool is something else! It’s a node-based AI image generator that gives designers crazy amounts of freedom and control. Imagine building your own image generation process like stacking LEGO bricks. Sounds fun, right?
As a visual designer, I had to dive in and see what’s possible. In this post, I’ll share my experience testing out ComfyUI, what makes it stand out, and how it compares to Midjourney (which works more like a gacha game where you type a prompt and hope for the best)，check out this link: 👉「Link」 . Also, if you want to dig deeper into PC vs. Mac performance in AI image generation, check out this link: 👉「Link」 .
Love experimenting with tech? ComfyUI, SDXL 1.0, 1.5, 3.5, and Flux Dev & Schnell are all open-source, offering endless possibilities!
ComfyUI’s 6 Coolest Features
Sure, ComfyUI can do the basic stuff like text-to-image, image-to-image, and using positive/negative prompts, but what really blew my mind were these:
Every generated image saves the workflow
If you import an image into ComfyUI, you can see the exact workflow used to create it. That means you can study, tweak, and learn from existing setups—super handy!
ControlNet is next level
ControlNet isn’t new (WebUI has it too), but in ComfyUI, it really shines! You can use a reference image to guide your AI generation, keeping the composition and layout consistent.
Stack multiple LoRA models in one workflow
Think of SDXL or Flux as your Art Director, setting the overall style. LoRA models? They’re like specialist designers, fine-tuning details without taking up too much storage. You can use three or more LoRA models at once in ComfyUI, giving you crazy customization power!
AI can help you write prompts!
You can connect large language models (LLMs) or import a CSV file of styles into your workflow. Say goodbye to writer’s block!
Seamless Photoshop integration
With plugins, you can link ComfyUI to Photoshop, making it part of your creative workflow.
Supports video & 3D models
You can import video models for animation. It even supports 3D modeling tools, meaning the possibilities go way beyond just 2D images.
The Power of Open-Source & Community Contributions
ComfyUI’s open-source nature is what makes it so powerful. The community is super active, and there are tons of third-party tools to expand what it can do.
Here are some must-follow creators:
📌 Bilibili: @秋葉aaaki – A true AI guru, offers a free SD package and keeps updating it!
📌 YouTuber: Pixaroma – Graphic designer who shares practical AI design tips.
📌 Bilibili: @Nenly – Great beginner-friendly ComfyUI tutorials.
Want more workflow ideas and models? Check out these:
📌 OpenArt – https://openart.ai/workflows/all
📌 Civitai – https://civitai.com/models
Future Wishlist: Making ComfyUI Even Better
✨ UI Improvements for Beginners
Right now, ComfyUI is powerful but not very beginner-friendly.
Could we get toolbars for key functions—like Photoshop has?
A more structured node layout would help new users connect components more easily.
🖥️ Cross-Platform Optimization
ComfyUI should run smoothly on PC, Linux, and Mac, but right now, performance varies.
Cloud-based deployment could be a game-changer. Maybe an official guide on setting it up?
☁️ My Failed Attempt at Running Flux on Google Colab
Flux is a set of high-end AI image models known for amazing quality and flexible styles. Since my Mac isn’t powerful enough, I tried running it on Google Colab.
Paid $10 for Google Colab to get extra cloud computing power.
Spent 10+ hours setting it up.
First few attempts worked fine, then I kept getting errors.
After multiple retries, I gave up and went back to SDXL. 😭
Moral of the story? AI art tools are evolving fast, but sometimes they can be a real pain to set up.
Final Thoughts
ComfyUI is an insanely powerful, highly customizable AI image tool, perfect for pro designers who need full control over their output. While the learning curve is steep, the freedom and precision you get make it totally worth it. If you love tinkering, customizing, and pushing AI to its limits, you have to try ComfyUI!
Some Personal Reflections
As of March 2025, we’re living in an AI boom. Every big company is racing to dominate the market. Just when I started getting comfortable with SDXL and Flux, boom, Flux 2.0 and SD 3.5 dropped. China’s Deepseek AI model also made headlines in February by bypassing Nvidia’s CUDA restrictions. Who knows what that means for AI art tools?
Technology is moving so fast, and as a designer, I feel both excited and nervous. Sometimes, it’s overwhelming. New tools pop up daily, and keeping up takes constant learning. But hey, I love the challenge! Even if it means struggling with code, I’ll do it.
Lastly, huge thanks to my fellow AI explorers:
Steven Dee – My partner & AI/3D design pro, who helped with the tech & resources for this post.
Tina Tsung – My awesome product designer friend, who inspired me to write this review.
You guys rock! ❤️
