Text-to-Image
Also called AI image generation, Image generation, Text-to-image model
AI that creates a new picture from a written description, usually using a diffusion model guided by the words in the prompt.
What is text-to-image?
Text-to-image is a kind of generative AI that turns a written prompt, such as "a watercolour of a lighthouse at dawn", into an original image. It's the technology behind AI image generators, and it's increasingly built into design tools, chat assistants and office apps.
How does it work?
Most text-to-image systems use a diffusion model. The model starts from random noise and removes it step by step, each time steering the picture closer to what the text describes. A text encoder translates the prompt into a form that guides each step. Some newer systems instead generate images directly inside a multimodal language model, which tends to follow complex instructions and render text in images more accurately.
Many tools also accept an image alongside the text, to edit it, change its style or keep a character consistent. Small add-ons such as LoRAs teach a model a specific style or subject.
What is it commonly confused with?
- Image search. The picture is newly generated, not found. It can still resemble existing works closely, especially when a prompt names a specific artist or character.
- Photo editing. Editing tools change an existing image; text-to-image creates one from nothing, though many tools now do both.
Why does it matter?
Text-to-image makes custom visuals fast and cheap for marketing, mock-ups, illustrations and ideas. Before you use the results commercially, check the tool's licence and its rules on real people and brands. Many tools mark their output with invisible watermarks or content credentials. Compare tools in Best AI image generators.
