Generating an image from a written description or prompt.
Text-to-image is a type of AI that turns a written description, called a prompt, into a generated image. You describe what you want in words and the model produces a matching picture.
It is powerful for ideation and one-off visuals, but text alone struggles to reproduce a specific person or product consistently. For brand work, text-to-image is usually combined with reference images and reusable elements to keep results on-brand.
Text-to-image starts from words only. Image-to-image starts from an existing image and transforms it, giving more control over the result.
Text alone usually cannot. Reproducing a specific product reliably needs reference images or saved elements.