Liquid Postman (Gemini Image)

A ComfyUI custom node for image generation using the Google Gemini API. Supports text prompts, image editing, PDF/DOCX file input, and search grounding.

Installation

ComfyUI Manager

Search for "Liquid Postman" in ComfyUI Manager and click Install.

CLI

comfy node install comfyui-liquidpostman

Manual

Clone this repository into your ComfyUI custom_nodes folder:

cd ComfyUI/custom_nodes
git clone https://github.com/majdsalim/ComfyUI-LiquidPostman.git
pip install requests Pillow

Setup

  1. Get a Gemini API key from Google AI Studio
  2. Add the Liquid Postman (Gemini Image) node to your workflow (found under google/gemini)
  3. Paste your API key into the api_key field

Inputs

Input Type Default Description
api_key STRING "" Your Gemini API key (required)
prompt STRING "" Text prompt for image generation
model STRING gemini-3.1-flash-image-preview Gemini model to use
seed INT 42 Random seed (-1 to disable)
aspect_ratio ENUM 1:1 Output aspect ratio (1:1, 2:3, 3:2, 3:4, 4:3, 9:16, 16:9, 21:9)
resolution ENUM 1K Output resolution (1K, 2K, 4K)
response_modalities ENUM IMAGE IMAGE or IMAGE+TEXT
system_prompt STRING (built-in) System instruction for the model
search_grounding BOOLEAN false Enable Google Search grounding
files STRING "" PDF/DOCX file paths or folder paths (comma or newline separated)
debug BOOLEAN false Print request/response details to console
images IMAGE (optional) Input images for editing or reference

Outputs

Output Type Description
image IMAGE Generated image(s) as a tensor batch
text STRING Text response from the model (when using IMAGE+TEXT mode)

Features

Image Generation

Enter a text prompt to generate images with Gemini.

Image Editing

Connect input images to the images pin. All images are uploaded via the Gemini Files API, so there is no file size limit.

File Input (PDF / DOCX)

Provide file paths in the files field. You can:

  • List individual file paths separated by commas or newlines
  • Provide a folder path to automatically include all .pdf and .docx files in that folder
C:\docs\report.pdf, C:\docs\brief.docx
C:\docs\project_files

Search Grounding

Enable search_grounding to let Gemini use Google Search for up-to-date information when generating images.

Debug Mode

Enable debug to print the full API request body and response metadata to the ComfyUI console.

License

MIT

S
Description
No description provided
Readme
36 KiB
Languages
Python 100%