Question
Can AI generate images from text?
Yes. Text-to-image models, such as Google's Nano Banana models and OpenAI's GPT Image models, turn a written description into a new image. The more specific you are about the subject, setting, light and style, the closer the first result tends to be to what you pictured.
A first result is a starting point, not a finished image. Google's guidance for its image models says not to expect a perfect image on the first attempt and to refine it with follow-up instructions, such as asking for warmer lighting.
Some things text-to-image still does less reliably. OpenAI lists precise text placement, keeping a recurring character or brand element consistent, and placing elements precisely in layout-sensitive compositions as current limitations of its image model.
In ImageQuarry you describe the image in plain language and it generates it. You then keep editing the result with words, and each change becomes a new version next to the earlier ones.
Sources: Google's developer documentation (Gemini image generation) and Google's Gemini best practices; OpenAI image generation guide (limitations); checked October 20, 2026.
Related questions
Last reviewed October 20, 2026 by ImageQuarry
Create an image. Edit it with words. Keep refining.
Describe your first image
Write what you want to see, get a first version, and keep editing it with words.
Start creatingAlready have an account? Sign in