AITiny can generate images from text prompts using AI image generation services. Generated images are saved to your Joomla Media Manager and can be used as article images, intro images, or anywhere else you use images on your site.
Image generation uses a separate AI service from the text generation features. You configure the image generation service in the Image generation tab of the plugin parameters. You need an API key for your chosen image generation service, which may be different from the API key used for text generation.
The image generation button appears next to Media fields in Joomla's article editor. This includes the intro image and full text image fields. The button uses the icon configured in the plugin parameters (by default, a laptop-code icon).
When you click the button, a modal dialog opens where you can enter your prompt and configure the generation options.
The image generation modal provides the following controls:
A text area where you describe the image you want to generate. Be as specific as possible about the subject, style, composition, and mood you want.
A button which uses the text AI service to automatically generate an image prompt based on the content of the article you are editing. This is useful when you want a contextually relevant image but are unsure what to ask for.
Select an image style preset. There are over 30 presets available, such as Photographic, Cinematic, Anime, Pixel Art, Watercolour, Studio Ghibli, and many more. The style prompt is appended to your prompt to guide the visual style of the result.
Select the output size: small, medium, or large. The exact pixel dimensions depend on the image generation service and model. Not all models support all sizes.
Select the aspect ratio for the generated image: 1:1 (square), 3:4, 4:3, 9:16, 16:9, and others. Not all aspect ratios are available for all models.
Once you have entered your prompt and selected your options, click to start generating the image. The generation may take several seconds depending on the service and model. You can click to cancel a generation in progress.
When the image is ready, it will be displayed in the modal. You can then:
Click to use the generated image. It will be saved to the configured media adapter and subdirectory, and the media field will be updated with the new image.
Click to generate a new image with the same prompt and settings.
Modify your prompt or settings and click again.
Click (or press Escape) to dismiss the modal without using the image.
When the Responsive Images option is enabled in the plugin parameters, AITiny automatically generates smaller variants of the accepted image at 512px, 256px, and 128px widths. These are saved alongside the original image and the appropriate srcset attribute is set on the image element. This ensures that visitors on smaller screens download a smaller, appropriately sized version of the image, improving page load times.
Be specific in your prompts. Instead of "a cat", try "a tabby cat sitting on a windowsill, looking outside at a rainy day, soft natural light".
Use the style presets to quickly achieve a specific visual look without having to describe it in your prompt.
Use the button as a starting point, then refine the generated prompt before sending.
If the result isn't what you wanted, click . Image generation is non-deterministic; the same prompt can produce quite different results on successive runs.
Different services excel at different types of images. For photorealistic images, OpenAI's GPT Image models and Google's Imagen tend to produce the best results. For artistic styles, Flux and Gemini can be excellent choices.