Happy Latent

A visual editor for generating videos with LTX 2.5 in ComfyUI. - Arrange scenes and adjust their duration on a timeline. - Guide each scene with a prompt and optional first and last frame images. - Choose generated audio, uploaded audio, or silence for each clip. - Upscale video resolution, increase frame rate, and refine the output. ## Install ### ComfyUI Manager 1. Open ComfyUI and click Manager. 2. Open Custom Nodes Manager and search for HappyLatent. 3. Select HappyLatent and click install. ### ComfyUI CLI ```sh comfy node install happylatent ``` ### Manual Navigate to your `custom_nodes` directory then clone the repository. ```sh git clone https://github.com/EricFillion/happylatent.git ``` Restart ComfyUI, then add the HappyLatentVideo node to your workflow. Generation requires the LTX 2.5 model, text encoder, video VAE, and audio VAE, which can be retrieved by running [workflows/starter/models.py](workflows/starter/models.py). You can find an example workflow in [workflows/starter/happylatent.json](workflows/starter/happylatent.json). ## Examples ### Image-to-Video

Generated Video	Node

### Multiple Scenes

Generated video	Node

### First and Last Frames

Generated Video	Node

Name	Type	Default	Description
images	IMAGE		—
vae	VAE		—
source_frames	INT	9–1201	—

HappyLatentTemporalSourceGuides

Inputs (3)

Outputs (1)