🚀 Run LTX-2.x with ComfyUI with provisioning — RunPod

int8 convrot

Workflow i2v

A streamlined and automated environment for running ComfyUI with LTX-2.x video models, optimized for use on RunPod

🔧 Features

• Automatic model and LoRA downloads via environment variables or lora-manager.
• Built-in authentication for:
  • ComfyUI
  • Code Server
  • Hugging Face API
  • CivitAI API
• Supports advanced workflows for video generation and enhancement using pre-installed custom nodes.
• Compatible with high-performance NVIDIA GPUs.

🧩 Template Deployment

Deployment.

• All available templates on runpod are tested on a L40S and RTX A5000.

Runpod templates

• 👉 One-click Deploy on RunPod LTX-2.5 i2v/t2v vi2v/vt2v dev INT8 ConvRot
• 👉 One-click Deploy on RunPod LTX-2.3 i2v/t2v vi2v/vt2v dev bf16/fp8

Documentation

• ⚙️ Start
• 📚 Tutorial
• ⚙️ Provisioning examples
• 🧩 RunPod environment profiles

🐳 Docker Images

Base Images

• PyTorch Runtime

• ComfyUI Runtime

Custom Image

docker pull ls250824/run-comfyui-ltx:<tag>

🛠️ Build & Push Docker Image (Optional)

Use the included Python script to build and push the Docker image.

Build Script: build_docker.py

Argument	Description	Default
--username	Your Docker Hub username	Current user
--tag	Custom image tag	Today's date
--latest	Also tag image as latest	Disabled

Example Usage

git clone https://github.com/jalberty2018/run-comfyui-ltx.git
cp ./run-comfyui-ltx/build_docker.py ..

export DOCKER_BUILDKIT=1
export COMPOSE_DOCKER_CLI_BUILD=1

python3 build_docker.py --username=<your_dockerhub_username> --tag=<custom_tag> --latest run-comfyui-ltx