Who is this course for?
ComfyUI is a powerful node-based interface built to drive the vast majority of 'AI' image processing tools. It is becoming the standard 'graphical user interface' of the 'machine learning' world.
Image professionals that want to integrate the latest 'AI' tools into their workflows should learn how to use ComfyUI.
If you are a compositor, matte painter, concept artist, designer, 3D generalist, technical artist, pipeline developer or simply a hobbyist, ComfyUI Fundamentals will get you started on the right track.
Go beyond 'prompt engineering'
Go beyond simple text prompts. Understand how 'AI' tools work under the hood and take control of your results.
'No spaghetti' guarantee
Clear and readable ComfyUI workflows built by a veteran of the VFX industry.
Learn by doing
Build your own ComfyUI workflows. Learn fundamental knowledge that you can build upon when new tools become available in the future.
Customer Reviews
Course Lessons
-
1Getting Started...
- Welcome to ComfyUI Fundamentals! (5:19) Free Preview
-
Safety and License information
-
Download Install Scripts, Workflows and Images.
-
2Installing ComfyUI
-
Setting Up ComfyUI on RunComfy (23:13)
-
Install on Windows 10/11 (23:49)
-
Install on macOS (23:16)
-
Install on Rocky Linux 9.x (25:25)
-
Install on Docker/Rocky Linux (37:25)
-
Install on Docker/Windows (41:39)
-
-
3Setting Up ComfyUI
-
Managing Custom Nodes (7:43)
-
Adjusting Preferences (7:45)
-
About Drag-and-Drop (11:07)
-
-
4Text-to-Image
-
Text to Image Generation (32:50)
-
KSampler Settings (19:35)
-
Installing Checkpoints (12:54)
-
-
5LORAs
-
Building a Slider LORA Workflow (17:26)
-
Building a Concept LORA Workflow (17:57)
-
-
6IPAdapter
-
Using Image Prompts with IPAdapter (11:48)
-
Building an IPAdpater Workflow (18:13)
-
-
7ControlNet
-
Exploring ControlNet Models (18:21)
-
Building a ControlNet Workflow (20:08)
-
-
9Inpainting
-
Using Inpaint Models (35:00)
-
Using Fooocus Inpaint (24:03)
-
-
10Upscaling
-
Using Upscale Models (16:42)
-
Building an Image-to-Image Upscale Workflow (19:00)
-
Building a Latent Upscale Workflow (18:49)
-
-
11Image Sequences
-
Building a Video-to-Video Workflow (31:01)
-
Building a Complex Video-to-Video Workflow (39:16)
-
-
12Student Survey
-
Help us improve!
-
What is included?
23+ Example Workflows
Exclusive Pixelfudger 'clear and readable' workflows. Ready to use right away in SD1.5 and SDXL versions.
6 Hours of Video Lessons
20 lessons for a total of 6 hours (approx.) of streaming video lessons. Watch and learn at your own pace.
Installation Guides
Install ComfyUI, models and required software easily with step-by-step video guides and exclusive Pixelfudger installation scripts for Windows, Mac, Linux and Docker.
Who is the instructor?
Xavier Bourque has been in the visual effects industry for more than 25 years and has been a professional instructor since 2017 at École NAD-UQAC in Montréal.
Example Lessons
-
Text-to-Image
-
Image-to-Image
-
Image Prompts (IPAdapter)
-
ControlNet
-
LoRAs
-
Inpainting
-
Upscaling
-
Image Sequence Processing
-
Combining multiple techniques
Frequently Asked Questions
-
No. Detailed video installation guides are included in this course to guide you through ComfyUI installation (including 'custom nodes' and 'AI' models.)
-
Check the minimum requirements below to see if your computer can run ComfyUI locally.
If your computer is not powerful enough to run ComfyUI locally, you can use RunComfy.com to run ComfyUI on a powerful machine in the cloud (Fees apply).
-
No. All the software used in this course is open source and available online for free.
-
The streaming video lessons have a total runtime of about 6 hours.
The installation guides are between 18 and 35 minutes each depending on the platform.
Including installation, watching videos and doing practical lessons, plan for about 12 hours to complete the course at a relaxed pace.
-
Please get in touch via the contact page to organize custom training for your team.
-
Une version française, en direct via Zoom, est offerte en collaboration avec l'École NAD-UQAC. Cliquez ici pour en savoir plus.
Minimum Requirements
Local installation
Windows 10 or 11: Nvidia GPU with 8GB of VRAM minimum (12GB+ recommended).
Rocky Linux 9.x: Nvidia GPU with 8GB of VRAM minimum (12GB+ recommended).
Apple Silicon Mac: 16GB+ of unified memory.
8GB of available disk space for Python, PyTorch and other modules.
35GB+ of available disk space for ComfyUI, custom nodes and models.
Broadband internet connexion to download software, models and stream video lessons.
Cloud Installation
Active pay-as-you-go or monthly account on RunComfy.com.
Up-to-date web browser (Chrome, Safari, Edge, etc...).
Stable internet connection.
Disclaimer
The best efforts have been made to ensure that the information in this course is accurate and up to date. However, you should be aware that ComfyUI is constantly evolving, and that some details may become outdated over time. Please report factual errors via our contact page.
Installing and using ComfyUI (and any associated software) is at your own risk.
Pixelfudger/Xavier Bourque is not responsible for any losses or damages that may arise from following this course or using any software demonstrated in this course including ComfyUI, 'custom nodes', or model files.
Additionally, it is the user's responsibility to ensure compliance with the licenses and terms of use for any models or software used in conjunction with ComfyUI, including those demonstrated in this course. Always review and adhere to the licensing terms of third-party models and software before commercial or personal use.