Home Creators Posts Import Register Favorites Logout

Downloads

Content

Hello Friends,

It’s here — the ComfyUI node that makes AI video generation fun again.

It’s easy to use, runs on as little as 12 GB of VRAM, and is incredibly powerful. If you want the full deep dive, be sure to check out my video tutorial.

Some YouTubers seem to believe that workflows they stitched together from the original creator's work in 10 minutes should be locked behind a $5–10 paywall. Yeah, I don’t see that…

That’s why below this post, you'll find the complete workflow and step-by-step installation instructions available completely for free.

Sure, I have paywalled content too, but that's content that:

a) I spent at least 50+ hours creating, refining, testing, and documenting, and

b) You typically won't find anywhere else (at least not with the same level of power, polish, or ease of use).

First of all, a huge shoutout to Jonathan Watkins for creating the repository that made all of this possible. 👏 👏 👏

https://github.com/WhatDreamsCost/WhatDreamsCost-ComfyUI

Without his work, none of this would exist.

That being said, supporters can take advantage of my installer software, which handles the entire setup process for you, including ComfyUI installation, model downloads, custom node installation, and all additional configuration steps.

Download from this post

For everyone else, here is the complete manual installation guide:

To follow along you will need:

🐍 Python Requirements

Python Version: 3.12

CUDA Version: 12.8

PyTorch Version: 2.8.0

📦 Git Repositories You Need

Main Repository (ComfyUI):

https://github.com/comfyanonymous/ComfyUI

Additional Repositories

rgthree-comfy

https://github.com/rgthree/rgthree-comfy

ComfyUI-GGUF

https://github.com/city96/ComfyUI-GGUF

ComfyUI_essentials

https://github.com/cubiq/ComfyUI_essentials

cg-use-everywhere

https://github.com/chrisgoringe/cg-use-everywhere

ComfyUI-KJNodes

https://github.com/kijai/ComfyUI-KJNodes

comfy-audio-duration

https://github.com/scofano/comfy-audio-duration

ComfyUI-Manager

https://github.com/Comfy-Org/ComfyUI-Manager

ComfyUI-MelBandRoFormer

https://github.com/kijai/ComfyUI-MelBandRoFormer

WhatDreamsCost-ComfyUI

https://github.com/WhatDreamsCost/WhatDreamsCost-ComfyUI

🎨 Models You Need

🖼 Diffusion Models

(Tip: Place in your models\diffusion_models folder)

📥 ltx-2-3-22b-V1.1-dev

VRAM Options (pick one that fits your VRAM):

12GB https://huggingface.co/QuantStack/LTX-2.3-GGUF/resolve/main/LTX-2.3-distilled-1.1/LTX-2.3-22B-distilled-1.1-Q2_K.gguf?download=true

16GB https://huggingface.co/QuantStack/LTX-2.3-GGUF/resolve/main/LTX-2.3-distilled-1.1/LTX-2.3-22B-distilled-1.1-Q3_K_S.gguf?download=true

24GB https://huggingface.co/QuantStack/LTX-2.3-GGUF/resolve/main/LTX-2.3-distilled-1.1/LTX-2.3-22B-distilled-1.1-Q5_K_M.gguf?download=true

32GB https://huggingface.co/QuantStack/LTX-2.3-GGUF/resolve/main/LTX-2.3-distilled-1.1/LTX-2.3-22B-distilled-1.1-Q8_0.gguf?download=true

📥 MelBandRoformer

VRAM Options (pick one that fits your VRAM):

12GB https://huggingface.co/Kijai/MelBandRoFormer_comfy/resolve/main/MelBandRoformer_fp16.safetensors?download=true

24GB https://huggingface.co/Kijai/MelBandRoFormer_comfy/resolve/main/MelBandRoformer_fp32.safetensors?download=true

📝 Text Encoders

(Tip: Place in your models\text_encoders folder)

📥 gemma_3_12B_it_fp4_mixed

https://huggingface.co/Comfy-Org/ltx-2/resolve/main/split_files/text_encoders/gemma_3_12B_it_fp4_mixed.safetensors?download=true

📥 ltx-2.3_text_projection_bf16

https://huggingface.co/Kijai/LTX2.3_comfy/resolve/main/text_encoders/ltx-2.3_text_projection_bf16.safetensors?download=true

🎭 VAEs

(Tip: Place in your models\vae folder)

📥 LTX23_audio_vae

https://huggingface.co/Kijai/LTX2.3_comfy/resolve/main/vae/LTX23_audio_vae_bf16.safetensors?download=true

📥 LTX23_video_vae

https://huggingface.co/Kijai/LTX2.3_comfy/resolve/main/vae/LTX23_video_vae_bf16.safetensors?download=true

🔍 Upscale Models

(Tip: Place in your models\latent_upscale_models folder)

📥 ltx-2.3-spatial-upscaler-2x-1.1

https://huggingface.co/Lightricks/LTX-2.3/resolve/main/ltx-2.3-spatial-upscaler-x2-1.1.safetensors?download=true

Loras Used:

Camera Controlshttps://civitai.com/models/2622189/camera-controls-ltx-23

Crisp enhancerhttps://civitai.com/models/2535622/ltx-23-enhancers

Files

The Most Powerful AI Video Generator Yet?

The new LTX Director for ComfyUI completely changes AI video generation. In this tutorial, I'll show you how to create longer, more cinematic videos using a powerful timeline editor, multiple prompts, multiple images, audio integration, camera controls, and prompt relays. You'll learn how to create image-to-video scenes, talking characters with audio, first-frame/last-frame animations without a first frame, multi-character story sequences, and much more. I'll also share the biggest mistake most creators make when generating AI videos and how to avoid it. Whether you're creating cinematic storytelling, character animations, AI shorts, or experimental projects, LTX Director is one of the most powerful additions to the LTX 2.3 ecosystem so far. Workflow and installation instructions are included. 📅 Book Your Personal 1-on-1 Session: 👉 Schedule Here: https://koalendar.com/e/1hr-1-on-1-stable-diffusion 🎁 Free Goodies & Essential Resources: • 🎯 Ultimate Beginner's Guide: https://www.patreon.com/posts/sneak-peek-alert-90799508 • 🛠️ Free Workflow Guide: https://www.patreon.com/posts/get-your-free-99183367 • 🎨 My Exclusive Style Collection: https://www.patreon.com/posts/my-collection-of-87325880 • 🚀 Support & Get Exclusive Content: 👉 Join Patreon Community: https://patreon.com/AIKnowledgeCentral 🎬 Popular Playlists: • 🤖 Invoke AI: Watch Here: https://www.youtube.com/watch?v=70HTvc-4gek&list=PL8HbYkSF5f9Ggts03f7ttPzFJyFiHgqoJ • ⚡ Flux: Watch Here: https://www.youtube.com/watch?v=70HTvc-4gek&list=PL8HbYkSF5f9Ggts03f7ttPzFJyFiHgqoJ • 🛠️ Forge: Watch Here: https://www.youtube.com/watch?v=70HTvc-4gek&list=PL8HbYkSF5f9Ggts03f7ttPzFJyFiHgqoJ • 🎨 ComfyUI: Watch Here: https://www.youtube.com/watch?v=70HTvc-4gek&list=PL8HbYkSF5f9Ggts03f7ttPzFJyFiHgqoJ • 🖥️ Automatic 1111: Watch Here: https://www.youtube.com/watch?v=70HTvc-4gek&list=PL8HbYkSF5f9Ggts03f7ttPzFJyFiHgqoJ • 📌 AI Tools: Watch Here: https://www.youtube.com/watch?v=70HTvc-4gek&list=PL8HbYkSF5f9Ggts03f7ttPzFJyFiHgqoJ 🌎Civitai: https://civitai.com/user/AIknowlege2go Chapters 00:00 Basics and Workflow 02:43 Image to Video 05:55 Creating Long Videos 07:40 Frist Frame - Last Frame without a First Frame 10:58 NSFW 12:35 Multi Image Input 14:00 Image + Custom Audio file

Comments

David Thielen

I've been doing other things for about 7 months and am now circling back to AI Images & Videos. Three questions: 1. What of your recent posts should I read (i.e. how far back should I read)? 2. What is the best model (and LORAs if needed) for NSFW Txt2Img and Txt2Vid. 3. Is there a way in Txt2Img and Txt2Vid to have it use the same character in subsequent images/videos? In other words create a character then continue to use it for subsequent generations. TIA

AI Knowledge Central

Regarding the best model. It depends. Depends. T2I would be ZImage Turbo, Ideogram 4, Krea 2 Turbo. Txt2Vid If you want Lipsync LTX 2.3 V1.1 with the Director is the way to go. Otherwise use Wan 2.2. For creating Consistent Characters In a longer video create Still Shots using either Qwen Image Edit or Flux 2 Klein Then Animate using LTX or Wan.

Kalle Blomquist

The Director Option is an awesome idea but it seems as if the I2V Option in Director is too much for my NVIDIA 4060... Even with 720p and 8 seconds it is difficult to get out videos that don't hurt my eyes... 1) Is there any negative prompting options that i missed? 2) The first round always works without problems but on the second round it always seems to crash. Any idea what could be the problem?