LTX 2.3: Start to End Frame Workflow build from scratch

Learn how to build a complete ComfyUI workflow from scratch for the LTX 2.3 video model. In this tutorial, we focus on generating video by locking in exact start and end frames, giving you much more control over the final animation. This setup uses the GGUF version of the Gemma 3 text encoder, which means the entire workflow is highly optimized and runs smoothly on an 8GB VRAM setup without crashing. I will walk you through downloading the correct models, installing the necessary custom plugins (like ComfyUI-GGUF and VideoHelperSuite), and wiring every node together step-by-step. What you will learn in this video: Downloading the LTX 2.3 model, VAE, and Text Projection files. Finding and installing working Gemma 3 GGUF files. Installing essential ComfyUI custom nodes (Manager, GGUF reader, VHS). Wiring up Guide Nodes to lock in exact start and end frames for precise video generation. Setting up the KSampler and rendering a smooth 24fps MP4. Resources & Setup Links: How to install ComfyUI:    • How to Install ComfyUI: Desktop App, Windo...   Download the Workflow (JSON): https://drive.google.com/file/d/1Gan3... Models & Files Needed: Gemma 3 GGUF: https://huggingface.co/mradermacher/g... LTX 2.3 Text Projection (found in the "text_encoders" folder): https://huggingface.co/Kijai/LTX2.3_c... LTX 2.3 VAE (found in the "vae" folder): https://huggingface.co/Kijai/LTX2.3_c... LTX 2.3 Unet GGUF: https://huggingface.co/QuantStack/LTX... Custom Nodes Used (Search in ComfyUI Manager): ComfyUI-GGUF (by city96): https://github.com/city96/ComfyUI-GGUF ComfyUI-VideoHelperSuite / VHS (by Kosinkadink): https://github.com/Kosinkadink/ComfyU...