Midjourney is currently the most popular image generation AI model in the world, with over 17.5 million users on Discord. Midjourney recently started testing the "alpha" version of its website, midjourney.com, which will allow users to generate images directly on the website without using Discord.

A passionate user recently discovered the Midjourney Alpha website with AI image generation features and shared screenshots and videos on X.

Currently, Midjourney has confirmed in its Discord announcement channel that it has "started testing the alpha version of the website with image creation capabilities," but the company notes that the alpha version is limited to "people who have generated over 10,000 images on Midjourney."

The company adds in the announcement, "Don't worry, we will open it up to more people in the next month!"

To let users know how many images they have generated and whether they have exceeded the threshold of 10,000 images, they can first log into Midjourney on Discord and then enter the slash command "/info" in any channel or directly to the Midjourney Discord bot for updated counts.

Appearance and Functionality

The website has a clean and minimalist interface with plenty of white space. It starts with an "Explore" tab that displays images publicly generated by other users. At the top, there is a blank rectangular text field labeled "Imagine," with a plus sign next to it, where eligible users can input prompts directly.

On the far right of this field is an icon consisting of lines and circles, representing a knob that users can click to launch a visual interface to adjust their prompt settings.

This is a significant departure from the current Discord version of Midjourney, which requires users to input various professional text commands in their actual prompts.

Therefore, with the web interface, users no longer need to input "--ar 4:3" to create a landscape image; they can simply click options in the prompt settings menu, and the system will automatically apply the setting without adding any visible text to the user's prompt.

In addition, using the buttons on the left navigation bar, users can click the "Photos" icon and see a reverse chronological view of all the images they have generated on the service since they started. These images are arranged in a grid by date. This view also shows the user's current image generation task.

Competition with Competitors

Overall, it is very similar to Visual Electric, a competing image generation web service based on the open-source Stable Diffusion model, which was released last week and targeted professional designers.

However, Midjourney's advantage lies in its many users' perception of high-quality image generation models. The Alpha website also includes a "Liked" tag, showcasing all the images that users have "liked" on Discord before.

The website also features a prompt search bar below and to the right of the image generation field, where users can enter a prompt keyword, such as "steampunk," and see all publicly generated images by other users using that prompt word.

In the bottom left corner, there is a "sun/moon" icon button that allows users to switch between light and dark modes. There is also a "like" button that allows users to participate in a crowd-ranking project for two images provided by the Midjourney team, which helps further improve the quality of their AI model.

However, one limitation is that the images generated using the web interface do not seem to reappear in the Midjourney Discord Bot chat. This means that the web interface is the only interface that retains all the images generated from both sources for users who switch back and forth.

Nevertheless, overall, this is an impressive alpha version release that will undoubtedly make Midjourney more accessible to a wider potential user base. However, for now, logging in with a Discord username is still required.