CivArchive
    Remove Metadata from Images, Video, Audio - ComfyUI-MetaWipe Workflow - v1.1
    NSFW
    Preview undefined

    Ignore base model - doesn't require a model, Civit just makes me choose one

    The workflow (config) and the nodes (archive) are both downloadable in the sidebar ->>>

    ComfyUI-MetaWipe

    URL: https://github.com/UltimateIntent/ComfyUI-MetaWipe/

    MetaWipe clears your sensitive data from images, videos, audio in ComfyUI:

    - MetaWipe (All-in-One): creates cleaned copies of media files with non-essential metadata removed.

    - MetaWipe Metadata Inspector: inspects media metadata before/after cleanup with a built-in per-file viewer.

    Why Use This

    • View and clear sensitive metadata before publishing to the web

    • Clean metadata from many sources

    • Clean single files or all folders/subfolders of mixed images, videos, audios

    • Keep final outputs organized while retaining your original copies

    • Prevent others from seeing your workflow and system details

    Installation

    1. Extract the archive zip (download in sidebar) into ComfyUI/custom_nodes/

    or git clone this repository into ComfyUI/custom_nodes/.

    git clone https://github.com/UltimateIntent/ComfyUI-MetaWipe/ 

    2. Restart ComfyUI.

    3. Add MetaWipe (All-in-One) and/or MetaWipe Metadata Inspector nodes to your ComfyUI workflow

    Runtime Requirements

    - Python packages: pillow, numpy

    - System dependency: ffmpeg available in PATH

    Node: MetaWipe (All-in-One)

    Creates cleaned media copies with stable sequential naming.

    Inputs

    - any_input (required): accepts media from:

    - file path strings (single or multiline)

    - directory paths

    - IMAGE VIDEO and AUDIO types

    - compatible wrapper objects

    - recursive (BOOLEAN, default false): if true, includes subfolders in metadata cleaning

    - output_subfolder (STRING, default %Y-%m-%d): target folder under ComfyUI output directory.

    - Supports strftime formatting %Y, %m, %d, etc.)

    - Supports environment variable expansion (Windows-style %VAR%)

    - output_filename_prefix (STRING, default clean_): prefix for saved cleaned files.

    Outputs

    - output_files (STRING): newline-separated list of created file paths.

    - any_out (Any): passthrough-oriented output for downstream chaining.

    Naming Rules

    - Uses incrementing sequence numbers 00000, 00001, ...).

    - Continues from highest existing sequence in the destination folder.

    - Includes original source filename in saved copy names.

    - Handles collisions with _(1), _(2), etc.

    Node: MetaWipe Metadata Inspector

    Inspects metadata and provides a selectable per-file JSON viewer in the node UI.

    Inputs

    - any_input (required): accepts the same input forms as MetaWipe.

    - recursive (BOOLEAN, default false): if true, includes subfolders in metadata viewer

    Outputs

    - metadata_text (STRING): readable metadata summary.

    - metadata_json (STRING): structured JSON metadata.

    - filepaths_out (STRING): list of inspected file paths.

    - any_out (Any): output of input for graph chaining

    Viewer Behavior

    - Dropdown selector lets you switch file-by-file.

    - Viewer panel displays pretty JSON for the selected item.

    Typical Workflows

    - Before/after verification

    - Any source -> Metadata Inspector (precleaned) -> MetaWipe -> Metadata Inspector (postclean verification)

    - Direct cleanup

    - Any source -> MetaWipe

    - Path-based chaining

    - MetaWipe.output_files -> Metadata Inspector.any_input

    Description

    expanded audio file support

    FAQ

    Comments (2)

    TheEventHorizonJul 26, 2026
    CivitAI

    is there a way to keep some meta data, like the prompts and models used so I can upload it to civit?

    ultimo_intento
    Author
    Jul 26, 2026

    it depends on what civit reads from an uploaded file to automatically tag the models, loras, prompts used. afaik, civit depends on reading the embedded workflow.

    since metawipe removes all embedded workflow data (as well as other metadata), you would need to design a workflow that loads the original 'dirty' file, dynamically generates a dummy workflow with the same prompts/models/loras/etc, and resaves file object, to obscure the original workflow while preserving the essential components in the metadata for automatic tagging

    Workflows
    ZImageTurbo

    Details

    Downloads
    84
    Platform
    CivitAI
    Platform Status
    Available
    Created
    6/2/2026
    Updated
    7/26/2026
    Deleted
    -

    Files

    removeMetadataFromImages_v11.zip

    Mirrors