Why I Switched from MidJourney: Discover the Cutting-Edge Alternative

We’ve talked a lot about Midjourney, and made no secret of our use of that service to generate art for articles. We love it! But one thing has really ground my gears since the service launched.

First, it’s the Discord-only interface. It’s like they had enough funding to build their GPU clusters and then ran out of money when it came time to build the interface, so they just hooked up a Discord bot and let it go at that. The Discord interface works fine but it’s a painful workflow.

There’s also no API (somewhat famously – people have been asking for it for more than a year) and to use Discord, you have to learn a lot of flags and commands. MJ did put a stub in their web UI where you could get your API keys…but you can’t do anything with them. Meanwhile, there are all kinds of third-party services hacking up APIs that on the backend talk to Discord…it’s pretty ugly.

So there’s a ton of demand for an API or at least a better interface, but MJ seems to be sitting on its hands.

Fortunately, others are not.

Leonardo is here.

Leonardo provides a web-based interface for image generation, a pleasant history of your projects to scroll through, and…well, it’s everything you wanted from Midjourney but couldn’t get.

For starters, everything in Leonardo is web-based. You don’t have to learn Discord, or configure your own Discord server and Midjourney bot. You can free all that mental RAM you’ve been using to remember obscure Midjourney flags and delete those PDFs with all those tricks. What Midjourney does by “command-line,” Leonardo does by GUI.

Leonardo offers a variety of models to explore. Cinematic, portrait, graphic design, anime, etc. are all accessible, with numerous sub-options and refinements. You can achieve similar results in Midjourney by using “in the style of,” but it’s more verbose and needs references.

There are numerous convenience features, such as a single click to remove backgrounds. I find the entire Leonardo interface quite user-friendly, as everywhere I want to copy or modify something, there’s a button or tooltip ready.

Did I mention there’s also an app? And an API? I wrote a quick Python script to generate a few hundred images and save each in a file with its seed, and it was effortless.

In my experience, Leonardo’s photorealistic outputs are slightly better than Midjourney’s, but the real comparison is about the interface. I doubt there is anything Leonardo can do that Midjourney can’t. But the experience is so much more enjoyable with Leonardo!

Leonardo has a limited free trial so give it a try. Maybe it’ll light a fire under the Midjourney dev team?