Midjourney is one of the most customizable and comprehensive services in terms of the illustration possibilities offered. Problem is, it is currently only used through the Discord messaging software. As competition rises, Midjourney can no longer rest on its laurels and is working on developing a solution accessible from a web browser. The recent arrival of Meta’s Imagine image generator as well as the integration of Dall-E into ChatGPT are probably triggers.
Midjourney Alpha: finally a graphical interface!
Midjourney could have simply replicated the current prompt entry method used on Discord in a web interface, but the reality is quite different. Developers are indeed seeking to make the generation of prompts more intuitive.
Rather than entering parameters on the command line in Discord, this new interface becomes graphical, equipped with scroll bars (slider) and buttons to select. For example, instead of writing “–ar 4:3” to indicate to Midjourney that a landscape format is desired, the user will click on the button corresponding to this ratio. In the same vein, certain customizations can be adjusted thanks to the appearance of a scroll bar, which allows to precisely adjust the intensity of a given parameter.
This new web version of Midjourney includes a tab that lists all the images that the user has previously liked in the Discord interface. A keyword searches section also allows you to search for any publicly generated image, as long as the prompt used contains the entered phrase.
The path taken by Midjourney could have a positive influence on the development of the company, by making this AI even more accessible to the general public. However, the circle of users currently eligible for Midjourney Alpha is still small. Midjourney has officially communicated on the subject via its Discord channel dedicated to announcements: the generation of images from the web interface is accessible to users with more than 10,000 images generated to their credit.