Interactively generate and modify images using OpenAI's gpt-image-1 model. Optionally upload an image or paste a URL, describe the changes you want, and chat to refine the result. A few sample images are loaded from config.json to try out the tool quickly.
- Image Selection
- Optionally upload an image or paste its URL.
- A set of sample images are provided as quick starting points.
- Prompt-Based Editing
- Describe how you want the image updated. If no image is chosen, a brand new one is generated.
- Iterative Chat
- Each generated image becomes the new base so you can keep refining it in subsequent prompts.
- Configure your OpenAI key and select an allowed base URL from LLM Foundry.
- Optionally upload an image or paste a URL and type a prompt describing your desired output.
- If an image is provided, it is sent to the
/images/editsendpoint. Otherwise/images/generationsis used. - The returned
b64_jsonis displayed and becomes the new base image. - Continue chatting with new prompts to iteratively improve the image.
Sample images and prompts are read from config.json on page load.