AirBrush offers an easy way to access multiple different image generation models. The fact that you can access a wide variety of models is the reason I give it 4 tacos.

Otherwise, as the product currently work, I'd give it 3. Although straigthforward: you enter a prompt, you select a mode, you get an image. In practice, it's very clunky. The prompt rarely gives a satisfying result the first time, and like most image generation tool, you have to wait a bit for the image. It would be useful if variations of the images could be offered from the get go - maybe even across multiple models so that we can have a feel for which model gets the prompt more. That would also save time. And then maybe a fast way to iterate over the images and prompts.

Other examples of the lack of polish: not all models support all the size. It's frustrating to request a large landscape and get back a squared image.

There are other tools like the prompt enhancer and the image variation tools. Great, but why do I have to go a separate page to use them, and lose my workflow? Having the prompt suggestions right when I'm entering them would make sense.

And to use the image variation tool, we have to upload a JPG. But what if I want variation of images just generated? Feels like I'm jumping through hoopses.

But if you want to be able to play with multiple models and don't mind a lot of waiting and iterations - it's great.