Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.comfy.org/llms.txt

Use this file to discover all available pages before exploring further.

This documentation was AI-generated. If you find any errors or have suggestions for improvement, please feel free to contribute! Edit on GitHub

SaveImageAdvanced

The SaveImageAdvanced node saves images to your ComfyUI output directory with advanced control over file format, bit depth, and color space. It supports saving as PNG or EXR files and can embed workflow metadata into the saved files.

Inputs

ParameterData TypeRequiredRangeDescription
imagesIMAGEYes-The images to save.
filename_prefixSTRINGYes-The prefix for the file to save. May include formatting tokens such as %date:yyyy-MM-dd% or %Empty Latent Image.width%. (default: “ComfyUI”)
formatCOMBOYes"png"
"exr"
The file format in which to save the image. Selecting a format reveals additional options for that format.
bit_depthCOMBOYes (conditional)For PNG: "8-bit"
"16-bit"
For EXR: "32-bit float"
The bit depth for the selected format. This parameter appears when a format is chosen. (default: “8-bit” for PNG, “32-bit float” for EXR)
input_color_spaceCOMBOYes (conditional)For PNG: "sRGB"
For EXR: "sRGB"
"HDR"
"linear"
Colorspace of the input tensor. For PNG, only sRGB is available. For EXR, the image is always written as scene-linear in the matching gamut. (default: “sRGB”)
Notes on Parameter Dependencies:
  • The bit_depth and input_color_space parameters are only available when a specific format is selected.
  • For PNG format, only “8-bit” and “16-bit” bit depths are available, and only “sRGB” color space.
  • For EXR format, only “32-bit float” bit depth is available, with “sRGB”, “HDR”, or “linear” color spaces.
  • The input_color_space parameter for EXR determines how the input tensor is interpreted:
    • "sRGB" — input is sRGB-encoded Rec.709; the inverse sRGB EOTF is applied.
    • "HDR" — input is HLG-encoded Rec.2020 (BT.2100); the inverse HLG OETF is applied to get scene-linear light.
    • "linear" — input is already scene-linear (Rec.709 primaries); written through unchanged. Use this for renderer/compositor output.

Outputs

Output NameData TypeDescription
imagesIMAGEA list of saved image results, each containing the filename, subfolder, and type (“output”). This output is used for UI display purposes.

Source fingerprint (SHA-256): 61e52bab8c28437cf648e4790823c15dbe0f758478635b0bd8b5cce785421fe5