Aitrepreneur
patreon
LTX-2 ULTRA WORKFLOW
🕑 Added 2026-01-17 19:57:37 +0000 UTCHey everyone! I’ve just released a new video on IMAGE TO VIDEO WITH SOUND KING, called
“LTX-2”. It’s an absolutely insane open source model that can generates videos with sound from either a simple text prompt, an image or even a sound file and can run on LESS THAN 8GB OF VRAM! This is one of the most fun model I have ever tried!
You can check out the video right here: https://youtu.be/KKdEgpA3rjw
I also created a 1-click installer for the Webui for my Patreon supporters for Local AND Runpod, that you can find right here: https://www.patreon.com/posts/148469807
Here the link to the manual install guide: https://pastebin.com/f7phSQWs
Down below you will also find the LTX-2_ULTRA_WORKFLOW.json for free for you to use!
As always, supporting me on Patreon allows me to keep creating helpful resources like this for the community. Thank you for your support - now go have some fun😉!
PS: there is a slight mistake in the workflow that always needs a sound file uploaded to work even if you don't want to but it's easy to solve, just dl that sound file and load it before doing any video generation: SOUND FILE
Comments
PTMarks
Great workflow. Works great. Do you have a version for FLF either for LTX-2 or LTX-2.3? Your WAN2.2 FirstLastFrame workflow is awesome. It would be great to have that for LTX. Thanks.
Alon Ashkenazi
?anyone knows?
Alon Ashkenazi
If I want to upgrade my comfyui setup with all needed models, etc. but I don't want to clean install every time you have somthing new. is there an easy way to use the installer scripts to update instead of to install clean?
Tiloz
got the same issue
Jackal32
Problem : LatentUpscaleModelLoader Error while deserializing header: incomplete metadata, file not fully covered
G
Hey guys anyone having or have solved this issue? LTXVLoopingSampler mat1 and mat2 shapes cannot be multiplied (1024x3840 and 15360x3840) TIPS: If you have any "Load CLIP" or "*CLIP Loader" nodes in your workflow connected to this sampler node make sure the correct file(s) and type is selected. I tried to change the clips and models from the unsloth to KJ and nothing works. gonna have to convert this into a fp8 version to make it work. If anyone knows how to solve this incompatibility please help.
Sergey Power
would love to see Frame To Frame added to this workflow at some point - this is something I've used a lot in your WAN 2.2 Workflow. Now I need to use another workflow for that which is slightly less convenient.
arman kezinch
I've tried the ComfyUI versions for LTX2 (Kijai's, Benji's, and yours), but the rendering was too slow and the quality was too bad. Then I switched to LTX2 with Wan2GP and it's day and night difference. I now make perfectly looking 10-second lip-sync videos in just 2 minutes (720p) or 6 minutes (1080p). I'm using an RTX 5090 + 96 GB RAM. I still don't understand why the quality and generation speed differ so much between ComfyUI and Wan2GP.
JT353
im trying to use the ultra workflow, and have spent 6hrs on chat gpt trying to figure out how to get the custom nodes to work. im using the runpod template listed. comfy is constantly failing to install these. is there any advise?
Ruti Bindehassan Gayle
can you update the workflow or at least give us alternatives?
Humhead
Working perfectly and a lot of fun to play with. Thank you!
Me Myself and I
No prompt enhancer in this workflow? LTX seems to like longer prompts, so enhancer can help improve a basic prompt. The default workflow that comes with LTX nodes has that if you want to look at it to add something similar for yourself or your audience to use.
Bob Winberry
im getting this error - any ideas?: Cannot execute because a node is missing the class_type property.: Node ID '#744:524'
Vincent Nestler
I like to follow along with you in the videos. Can you please publish the prompt you use when you share. It would be very helpful!
