---
title: "Cogvideox Vae in ComfyUI"
description: "Cogvideox Vae is a VAE (Variational Autoencoder) that encodes and decodes latent representations."
canonical: https://comfy.org/p/supported-models/cogvideox-vae
lang: en
index: https://comfy.org/llms.txt
---

# Cogvideox Vae in ComfyUI

Cogvideox Vae is a VAE (Variational Autoencoder) that encodes and decodes latent representations. You can run it locally in ComfyUI with full control over every parameter, or access it through Comfy Cloud. ComfyUI's node-based workflow editor lets you connect Cogvideox Vae with ControlNets, LoRAs, upscalers, and custom nodes to build any pipeline you need. There is 1 community workflow template using Cogvideox Vae on Comfy Workflows, ready to load and customize.

## Facts

- Type: VAE
- Pricing: free to run locally on your own hardware; on Comfy Cloud it meters as GPU time at the published Cloud GPU rate: https://docs.comfy.org/tutorials/partner-nodes/pricing#cloud-gpu
- Community workflow templates: 1
- Weights: https://huggingface.co/Comfy-Org/void-model/resolve/main/vae/cogvideox_vae.safetensors
- Tutorial: https://docs.comfy.org/tutorials/utility/void-video-inpainting

## How to run Cogvideox Vae

1. Locally in ComfyUI — open source, free on your own hardware: https://comfy.org/download
2. On Comfy Cloud — hosted GPUs, every parameter still exposed: https://cloud.comfy.org
3. Start from a community workflow template and adjust it node by node: https://www.comfy.org/workflows
4. From your own application with the Comfy SDKs (Python and TypeScript, beta): https://docs.comfy.org/development/api-development/sdks.md
5. From a terminal or a coding agent with Comfy CLI, which runs workflows on Comfy Cloud or a local ComfyUI: https://docs.comfy.org/agent-tools/cli.md
6. From Claude Code, Cursor, or Codex over Comfy MCP at https://cloud.comfy.org/mcp: https://docs.comfy.org/agent-tools/mcp.md

This page as HTML: https://comfy.org/p/supported-models/cogvideox-vae
Full model catalog: https://comfy.org/p/supported-models/llms.txt