Google Gemini 2.0 Flash Image Generation Toolkit for Open WebUI

Author: Justin Miller Support This Project: Buy Me A Coffee

Updated - 11/12/2025

Gemini 2.0 Flash Preview Image Generation in the Gemini API will be discontinued on November 12, 2025. You can still use the tool and use their other image generation models, for example, gemini-2.5-flash-image, but as of today, they don't have any models that offer image generation in the free tier.

Updated - 7/20/25

Gemini has updated the 2.0 image generation to "Preview" from "Experimental."

Overview

This toolkit integrates Google's powerful (and experimental) gemini-2.0-flash-preview-image-generation model directly into Open WebUI, allowing you to generate images based on text prompts within your chat interface.

Leveraging the Gemini API, this tool can generate images and potentially accompanying text in response to your prompts. The results (both image and text, if generated) are displayed directly in the Open WebUI chat.

Note: This tool utilizes the gemini-2.0-flash-preview-image-generation model, which is currently experimental [0, 1]. Google documentation notes that for this specific model, the response modalities must include both "Text" and "Image" [1]. This toolkit handles that configuration automatically.

Features

Image Generation: Create images directly from text prompts using Google's Gemini 2.0 Flash model.
Text + Image Output: The model can return both generated text and an image in a single response. This tool handles displaying both in the UI.
Configurable:
- Set your Google AI API Key via the Open WebUI tool settings.
- Choose the output image format (PNG, GIF, JPEG).
- (Advanced) Change the specific Google AI model name if needed (though it defaults to the required experimental model).
Seamless Integration: Works within the Open WebUI tools framework, sending status updates and results directly to the chat interface.

Requirements

Open WebUI: Version 0.5.3 or higher
Python Packages:
- google-genai
- google-generativeai
- Pillow
Google AI API Key: You need a valid API key from Google AI Studio or Google Cloud.

Installation

Ensure you have the required Python packages installed in your Open WebUI environment:
bash
```
pip install google-genai google-generativeai Pillow
```
Place the tool's Python file (.py) into your Open WebUI tools directory as a new tool, or grab it from the Open WebUI Community
Restart Open WebUI.

Configuration

Navigate to the Tools settings within Open WebUI.
Find the "Google Gemini 2.0 Flash Experimental Image Generation" tool.
Enter your Google AI API Key in the api_key field. This is required for the tool to function.
(Optional) Change the desired image_format (default is png).
(Optional) Change the model_name if Google updates or changes the experimental model endpoint.

Usage

Once installed and configured, you can invoke the tool within Open WebUI like any other tool.

Example prompt:

Generate an image of a futuristic cityscape at sunset

The tool will:

Show a "Generating an image..." status message.
Call the Google Gemini API with your prompt.
If the API returns text, it will be displayed in the chat.
If the API returns an image, it will be uploaded via Open WebUI's image handling and displayed in the chat.
Provide a final status update (success or error).

Important Notes

Experimental Model: The gemini-2.0-flash-preview-image-generation model is explicitly marked as experimental by Google. Its availability or behavior may change.
API Key: Keep your Google AI API key secure. Do not share it publicly.
Safety Settings: The underlying Google API may have safety filters that prevent generation for certain prompts. The tool will attempt to relay any error messages back to the user.

Enjoy generating images directly within Open WebUI!