ComfyUI Tutorial #1 — Installation Guide for Beginners (Portable Version)
1. What is ComfyUI?
ComfyUI is a powerful, modular interface for Stable Diffusion that lets you build your own image generation workflows using visual nodes instead of text commands.
Think of it like LEGO blocks for AI art — extremely flexible, visual, and fast once you get used to it.
2. Why Use the Portable Version?
If you’re new to AI or just want something that works right away, the portable version is the best choice.
No need to install Python or Git
Works directly out of the box
Fully self-contained in one folder
Easy to back up or move between drives
3. Downloading ComfyUI (Portable Build)
3.1 Step-by-step:
Go to the official ComfyUI GitHub page:
- https://github.com/comfyanonymous/ComfyUIScroll down to “Releases” on the right sidebar
or directly open:
- https://github.com/comfyanonymous/ComfyUI/releasesDownload the latest ComfyUI Windows Portable ZIP file
(Example:ComfyUI_windows_portable_nightly.zip)Extract the
.zipfile somewhere simple, like:C:\ComfyUI\Avoid installing it inside “Program Files” or a protected folder — this can cause permission issues.
4. Launching ComfyUI
Once extracted:
Open the folder
ComfyUI_windows_portableDouble-click
run_nvidia_gpu.batif you have an NVIDIA GPU
(orrun_cpu.batif you don’t have one — slower but works)Wait for the window to finish loading — you’ll see lines of text appear (it’s downloading models).
When it says something like:
Running on http://127.0.0.1:8188
→ Open that link in your browser.
🎉 You’re now in the ComfyUI interface!
5. First-Time Setup
Once ComfyUI is running, you’ll need to add your Stable Diffusion models and VAEs.
5.1 Folder locations:
Models →
ComfyUI_windows_portable\models\checkpointsVAE →
ComfyUI_windows_portable\models\vaeLoRA →
ComfyUI_windows_portable\models\loraEmbeddings →
ComfyUI_windows_portable\embeddingsControlNet →
ComfyUI_windows_portable\models\controlnet
5.2 Download your base model
You can find many models (like SD1.5, SDXL, DreamShaper, etc.) on:
Once downloaded, drop the .safetensors or .ckpt file into your checkpoints folder.
6. Test if Everything Works
In ComfyUI, click Load Default Workflow
Click Queue Prompt (bottom right)
Wait for it to render
→ If an image appears in the preview, your setup works perfectly! ✅
7. Installing Custom Nodes (Optional but Useful)
Custom nodes add new tools and features like ControlNet, IPAdapter, or LoRA Managers.
7.1 Installation:
Go to your ComfyUI folder
Open:
ComfyUI_windows_portable\custom_nodes
Create a new folder for your node (example:
ComfyUI-Lora-Manager)Go to the node’s GitHub page (e.g. ComfyUI-Lora-Manager)
Click “Code → Download ZIP”, extract it into your new folder
Restart ComfyUI
💡 You can also install nodes directly with ComfyUI Manager (a built-in tool in newer versions).
8. Keeping Things Organized
If you plan to train or use multiple models, it’s smart to create subfolders:
models/
├── checkpoints/
│ ├── SDXL/
│ ├── SD1.5/
├── lora/
│ ├── characters/
│ ├── styles/
This helps ComfyUI load faster and keeps everything clean.
9. Common Problems & Fixes
9.1 Missing Node Type
Some nodes require a newer version of ComfyUI.
→ Update by re-downloading the latest portable build.
9.2 Models not showing
Make sure your model files are in the correct checkpoints folder.
9.3 CUDA / GPU error
Update your NVIDIA drivers and ensure you have at least 6GB VRAM for SDXL.
10. Optional Enhancements
ComfyUI Manager → Easy node installation
Impact Pack → Extra image filters
Wan Video Node → Animate your generations
Lora Manager → Manage multiple LoRAs in one click
11. Final Tips
Always keep one backup copy of your
ComfyUIfolder — it contains all your settings.Use seed locking if you want repeatable results.
Explore the Workflow Examples on CivitAI to learn how different setups work.
12. Conclusion
That’s it — you now have a fully functional ComfyUI portable setup!
You can start generating, building workflows, or even training your own LoRAs without touching a single command line.
“Once you start dragging nodes, you’ll never go back to text prompts.” 😄
13. Join the Discord!
Come hang out, share your AI art, give feedback, ask questions, or just vibe with the community.
It’s also the best place for troubleshooting, commissions, and sneak peeks of upcoming work.
.jpeg)

