TUTORIAL3D.IT · AI AND COMFYUI
AI and ComfyUI: all our tutorials
Here you’ll find all our articles and tutorials on generative artificial intelligence, explained by people who work with graphics, video and 3D every day. At the heart of it all is ComfyUI, the free program that lets you run on your own computer the models that create images, video and audio: Stable Diffusion, Flux, Qwen-Image, Z-Image, Wan, LTX and many more.
If you’re starting from scratch, begin with the introductory guide, then install ComfyUI and find out which graphics card you need. In the articles you’ll find model guides with licenses explained (so you know what you can also use for work), comparisons between local AI and online services, our free workflows and, every week, our editorial AI of the week. Every tutorial has its own video on our YouTube channel.
🧩 Free workflows
Three ready-to-use workflows for ComfyUI, with step-by-step instructions: Stable Diffusion 1.5, SDXL and Flux Schnell. They work even with little video memory.
▶️ Our YouTube channel
Every tutorial has its own video: installation, models, workflows and news. Subscribe so you don’t miss the next episodes on images, video and audio with AI.
📰 AI of the week
Every Thursday, the most important news from the world of artificial intelligence, explained from the point of view of people who create.
All our articles on AI and ComfyUI
Newest first: tutorials, model guides, comparisons and news. New articles show up here automatically.
Blender 5.3: DLSS in the Viewport, Ripple Editing and Vulkan
Blender 5.3 brings NVIDIA’s DLSS denoiser to the Cycles viewport (driver 620+), ripple editing and OpenTimelineIO in the Video Sequencer, a faster compositor and Vulkan by default. Here’s what changes and whether it runs on an RTX 3060.
ComfyUI 0.42: MOSS-Image Goes Stable, with Fast LoRAs and Lightweight GGUFs
ComfyUI 0.42 brings MOSS-Image to the stable release, with new 4- and 8-step LoRAs, Kromoss and 3–5 GB GGUFs. ostris opens the door to editing and Hi3D announces its open release.
Cinema 4D 2027 with Scatter and Liquid FX, plus AI Relighting in After Effects
Maxon releases Cinema 4D 2027 with Scatter and Liquid FX, Red Giant brings AI relighting to After Effects, LightVAE speeds up Wan and H3 videos, and ComfyUI’s LoRA stack nodes are waiting for the new frontend.
NVIDIA Nodes for Upscaling, Slow Motion and HDR in ComfyUI, Plus Fish Audio Drama 3
NVIDIA VFX nodes for upscaling, slow motion and HDR in ComfyUI, the stable ComfyUI 0.40.0 release, RTX Spark laptops with Adobe in beta and Fish Audio Drama 3 with open weights.
72 Shots of the Same Object with Qwen-Image 2.1, Plus the XPS 16 with RTX Spark
An Apache 2.0 LoRA for Qwen-Image 2.1 generates 72 shots of the same object, Dell opens pre-orders for the XPS 16 with RTX Spark, ComfyUI 0.39.2 brings new partner nodes, and SGF+ research tackles long videos.
Kandinsky 6.0: MIT-Licensed Video with Audio, and Nano Banana 2.1 Costs Half
Kandinsky 6.0 brings MIT-licensed video with audio and an official ComfyUI extension, Nano Banana 2.1 halves the API price of images, ComfyUI adds the CameraAngle node and Prism aims for 2K.
Pika H3 Flex in ComfyUI: Video and Audio in 6 Steps (and the License)
Pika H3 Flex brings MiniMax H3 down to 6 steps, with video and audio together and three ComfyUI templates. VRAM, files to download and the license catch if you work from Italy.
ComfyUI 0.39 with 16-bit EXR, Fizgig v7 and LoRAlab for LoRA Training
ComfyUI 0.39.0 brings 16-bit EXRs and an offline mode, Fizgig v7 does full fine-tuning from 8–12 GB, LoRAlab bundles nine LoRA trainers and VNCCS 3.2 creates consistent characters.
Wan and MiniMax H3 in a Few Steps with LongLive-Plug, plus Splats in Blender
NVIDIA brings Wan and MiniMax H3 down to a few steps with the LongLive-Plug LoRAs, Blender 5.3 shows Gaussian splats and DLSS in the viewport, and ComfyUI saves 16-bit EXR.
Free Higgsfield Alternative: 120 Apps in ComfyUI with Freefield
Freefield rebuilds Higgsfield’s apps as ComfyUI workflows with open-weight models: 120 apps for image, video and audio. VRAM, licenses and how to install it.
ComfyUI 0.38, LTX-2.5 in 4K and 8K and OpenAI’s DevDay Announcements
ComfyUI 0.38.0 is the new stable release and makes SeedVR2 lighter on VRAM, but watch out for a Qwen-Image 2.1 bug. LTX-2.5 reaches 4K and 8K, and OpenAI bets on agents at DevDay.
Vlo Connects Video Editing to ComfyUI and AnyAngle Changes the Framing
Vlo 0.3 connects video editing to ComfyUI, the Qwen-Image 2.1 ControlNet Union lands in ComfyUI’s code, AnyAngle changes the framing with Blender and LanPaint 2.2 retouches transparency too.
WORKFLOW GRATUITI
Scarica i workflow. Gratis.
I workflow ComfyUI pronti all’uso che usiamo nei nostri tutorial, con le note per installarli. Lasciaci la tua email e te li mandiamo subito.