A new artificial intelligence startup is challenging more established rivals by solving a vexing problem: letting users create an image with text you can actually read.

Like its peers, Ideogram can whip up startlingly realistic images from short text prompts in a matter of seconds. But Ideogram, a Toronto-based startup that launched in August, can also go a step further and render text within those images. It can generate an image of a protester holding a legible sign or a cute cat in a t-shirt that clearly says, “Ask me about my AI startup.” Resolving what may seem like a niche technical issue has broad implications for the industry. When asked to render words in pictures, other popular AI image generators such as Midjourney, OpenAI’s Dall-E 2 and Stability AI’s Stable Diffusion often show nonsense.

Up Next

Ideogram Produces Text in AI Images That You Can Actually Read