Image ProcessingPartial Capture

VAE Encode

class_type: VAEEncode

Source repo

Encodes a pixel image into a latent representation using a VAE. Used in img2img workflows where an existing image serves as the starting point for generation instead of random noise.

What It Does

VAEEncode is the inverse of VAEDecode—it converts a pixel image into the latent space representation that the diffusion model operates on. This is the entry point for img2img workflows, inpainting, and any technique that starts from an existing image rather than pure noise.

The provenance gap with VAEEncode is significant: the input image content is not embedded in the output PNG. The workflow JSON records the connection to a LoadImage node (which has a filename), but the actual image data is not preserved in the output file.

Inputs

pixelsIMAGE

Input pixel image to encode.

vaeVAE

VAE model for encoding.

Outputs

LATENTLATENT

Encoded latent representation.

What Numonic Captures

  • Connection to source image node in workflow graph

Known Gaps

  • Input image content — not embedded in output
  • Input image provenance (was it AI-generated? From where?)

Related Nodes

VAE DecodeImage Processing
Load ImageI/O
KSamplerSampling
Back to ComfyUI Nodes

Capture ComfyUI metadata automatically

Numonic extracts workflow metadata from every ComfyUI generation — models, samplers, seeds, prompts, and custom nodes. Track provenance, maintain compliance, and never lose a workflow.

Return to the Guide