Mann-E (Persian: مانی) is a state-of-the-art AI model, developed and trained by Muhammadreza Haghiri in late 2022 and then released as a universal accessible platform in April 2023.
Mann-E was originally based on Stable Diffusion but later we've developed our own diffusion models by standing on shoulders of the giants in the industry such as Stability AI and Meta. Our model in a lot of benchmarks outperformed Dall-E 3, Midjourney, Leonardo and even Ideogram!
Mann-E can understand prompts in a variety of languages. So you just need to input your prompts in your own language. For example if you speak Persian, you can write in Persian. If you speak Tamil, just write in Tamil. Our model can understand your language very well!
If you check 'Refine with AI', our language model will refine and revise your prompt and make it more understandable for the image generation model. Using this method, you easily can make images ten times better!
The big problem with most of image generators is that they cannot generate transparent images. We've solved this problem. You easily can generate transparent images and use them in your design projects!
