Gemini 2.0 in ComfyUI

Introduction

Google's latest multimodal large language model Gemini 2.0 Flash can generate and edit images through continuous dialogue, marking an important step forward in the future trend of LLM and intelligent human-computer interaction; a creator optimized its integration with ComfyUI (based on CY-CHENYUE's original project), fixed image conversion and API key security issues, allowing users to seamlessly explore Gemini 2.0's graphic capabilities in ComfyUI, and easily combine it with plug-ins such as HD Zoom. More graphic features will be added in the future to further enhance the creative task experience.

https://github.com/CY-CHENYUE/ComfyUI-Gemini-API

Recommended machine：Large-Pro

Workflow Overview

How to use this workflow

This workflow is divided into two parts. You need to call the Google API first, and then run the workflow after setting up the API

Part 1: Get Google API

Step 1: Visit Google AI Studio

https://aistudio.google.com/apikey?hl=zh-cn

If you cannot access it through the link, please refer to the link provided by the author's Readme in the Github link mentioned in the above introduction

Step 2: Apply for API

Click the boxed area to apply for your exclusive API, which will then be displayed in the display bar below

Part 2: Img2Img

Step 1: Upload reference image

Step 2: Input the Prompt

Supports Chinese and English languages, and Chinese comprehension is better than English

You can use this plug-in to do text2image (requires bypass load image), image editing, watermark removal, line drawing, e-commerce promotional images, image background change.

If you want to keep the character consistent, please state it clearly in the prompt.

Step 3: Add the API just generated

This workflow provides a temporary API for MimicPC designer testing. In order to avoid subsequent failure, it is best to follow the process of Part 1.

Step 4: Get Image

Step 5: View the Log

The custom_node author only supports Chinese, and does not support other languages ​​for viewing.