Patched a request forgery (SSRF) issue in the Python code that was keeping the node from being published in the ComfyUI Manager.
This node pack is still functional but slightly superseded by https://github.com/CovertBannana/ComfyUI-ImagePlus, which is different in the fact that it is a more Unified node set, sliming the number of nodes down and also add some additional functionality like local image upload.
- Fixed stale metadata collisions between CivitAI-loaded images and locally loaded image metadata.
- Fixed Image Preview state clearing so local image loads no longer retain previous CivitAI preview backend data.
- Fixed duplicate browser tab opening from the Image Preview node's Open Image button.
- Improved local image metadata handling to avoid pulling stale CivitAI prompt data from embedded workflows.
If updating from an earlier version, restart ComfyUI and hard-refresh the browser after updating the custom node.
###Fixed an issue where text was not persistent when changing tabs.
####Local Image Info node
v0.3.0 adds a new CivitAI Local Image Info node for working with images, containing metadata, stored locally.
You can now:
- Upload or drag/drop a local image directly onto the node.
- Extract positive and negative prompts from:
- ComfyUI PNG metadata
- ComfyUI workflow metadata
- CivitAI ImagePlus gallery metadata embedded in saved workflows
- Automatic1111/Forge-style metadata
- JPEG EXIF metadata where available
- Automatically send extracted prompts directly to the CivitAI Prompt Editor node.
- Send model, sampler, seed, VAE, CLIP, checkpoint, UNET, and generation settings to the CivitAI Info Display node.
- Preview the dropped/uploaded local image in the CivitAI Image Preview node.
- Open the preview image directly from the Image Preview node.
The CivitAI Image Preview node now includes an Open Image button.
- For CivitAI images, this opens the proxied ComfyUI preview image.
- For local uploads, this opens the local browser preview image.
- The existing Copy CivitAI Page URL button remains available for CivitAI-sourced images.
- Fixed: Corrected metadata parsing to ensure prompts extract properly after recent Civitai metadata restructuring.
- Added: Enhanced the Info block to cleanly parse and display the Sampler, Scheduler, Steps, Model, and CFG scales where available.
- Prevented: Stopped LoRA, embedding, and textual inversion names from accidentally overwriting the primary model checkpoint name.
Browse CivitAI images in ComfyUI, preview protected images, edit prompts before generating, and pass selected images directly into img2img workflows.
Tested on current stable ComfyUI releases (verified on v0.19.x at time of writing).
ComfyUI_CivitAI_ImagePlus adds a CivitAI gallery browser to ComfyUI, along with helper nodes for:
- Image preview with an img2img connector
- Direct prompt editing
- Quick generation info display
Nodes are populated instantly via canvas actions without needing to queue the workflow first.
- Open a full-screen CivitAI gallery from inside ComfyUI
- Browse images using:
- Most Reactions
- Most Comments
- Newest
- Pick an image and instantly:
- preview it
- preview protected images if you have a valid CivitAI API key
- load its prompts into an editor
- see a handy info panel
- Fetch by URL using a CivitAI page link
- Use the selected preview image as input for img2img workflows
CivitAI now uses separate public-facing domains:
https://civitai.comhttps://civitai.red
This plugin supports both for:
- Fetch URL input
- copied page links
- protected preview routing
The gallery includes a Page Links selector so you can choose:
- Auto
- Civitai.com
- Civitai.red
If you paste a CivitAI URL directly, the plugin will try to preserve that domain automatically.
Example workflows:
- workflows/Gallery_text_to_image.json
- workflows/Gallery_image_to_image.json
- workflows/Gallery_Z-image_Turbo.json
Note:
- The text-to-image and image-to-image workflows use ComfyUI core nodes plus this node pack.
- The Z-image Turbo workflow uses additional custom nodes.
- Open ComfyUI → Manager
- Search for CivitAI ImagePlus
- Install the node pack
- Restart ComfyUI
- Go to your ComfyUI
custom_nodesfolder:ComfyUI/custom_nodes/
- Clone this repo:
git clone https://github.com/CovertBannana/ComfyUI_CivitAI_ImagePlus.git- Restart ComfyUI
- If you are having trouble with Git, click the Code button on GitHub and download the repository as a ZIP
- Extract the folder into your ComfyUI
custom_nodesdirectory - Restart ComfyUI
To preview and fetch protected content, you need a CivitAI API key.
- Create the file if it does not exist:
custom_nodes/ComfyUI_CivitAI_ImagePlus/api_key.txt- Put your key inside like this:
CIVITAI_API_KEY=YOUR_CIVITAI_API_KEY_HERE✅ Restart ComfyUI after adding or changing the key.
Protected image previews are handled server-side.
- Your API key stays on your machine in
api_key.txt - The browser UI does not need direct access to your key
Open the gallery and select images.
- Open Gallery → full-screen browsing
- Fetch URL + Fetch → paste a CivitAI URL and load the image
Tip: you can use either:
- a post URL like
https://civitai.com/posts/<id>orhttps://civitai.red/posts/<id> - an image URL like
https://civitai.com/images/<id>orhttps://civitai.red/images/<id>
- Shows the selected preview image directly in the node
- Includes Copy CivitAI Page URL button
- Output:
IMAGE
This makes it ideal for img2img workflows.
Lets you edit prompts before generating.
- Shows Positive and Negative previews directly on the node
- Bottom buttons:
- Edit Positive
- Edit Negative
- Restore → revert to prompts from the last selection
- Clear → wipe both prompts
Outputs:
Positive(STRING)Negative(STRING)
A small info preview panel that updates instantly when you select an image.
Upload or drag/drop a local AI-generated image and automatically extract available metadata.
Supported metadata includes:
- ComfyUI prompt/workflow metadata
- Automatic1111/Forge-style parameters
- JPEG EXIF
ImageDescription/UserCommentwhere available
The node updates:
- CivitAI Prompt Editor
- CivitAI Info Display
- CivitAI Image Preview
- Select an image in CivitAI Gallery
- Edit prompts in CivitAI Prompt Editor
- Connect Prompt Editor outputs to your text / CLIP encode nodes
- Generate
- Select an image in CivitAI Gallery
- Confirm it in CivitAI Image Preview
- Use Preview output as your img2img init image
- Edit prompts in Prompt Editor
- Restart ComfyUI
- Hard refresh your browser (
Ctrl+F5)
- Confirm
api_key.txtexists and is formatted correctly:
CIVITAI_API_KEY=...- Restart ComfyUI
- Prefer using a post URL (
/posts/<id>) when possible - Some content may still be restricted by CivitAI account, domain, or region visibility rules
Use the gallery’s Page Links selector:
- Auto
- Civitai.com
- Civitai.red
If you fetch by URL, the plugin will try to preserve that domain automatically.
- Tested on current stable ComfyUI builds
- Uses standard ComfyUI frontend extension registration
- If you are testing against nightly/master builds, upstream frontend changes may land before a stable release
MIT License (see LICENSE).
Inspired by earlier community work exploring similar workflows and UI ideas.