Midjourney: How to Process Photos for Free as Good as in Lensa

A few words about the neural network

The Midjourney team is an independent research lab working on expanding humanity’s creative abilities.

Their AI creates images from text descriptions, similar to DALL-E from the OpenAI and Stable Diffusion open-source.
The network is currently beta testing so that everyone can try it.

To do this, you must provide the plot of the picture you want to create, but only in English. Everything is limited only by your imagination, but note that you have only 25 free attempts. After that, you can improve the result: enhance the quality and download it yourself.

The result is a full-fledged picture.

AI works like a discord bot: you make up a story, send it to the chat, and receive a generated picture a few minutes later.

Instruction for processing

• Install Discord, register, or log into your profile. It is necessary, as the neural network works through the discord bot.
• Go to — midjourney.com, click “join the beta,” and accept the invitation.

• Give the AI all the necessary rights and permissions it asks for 🙂
• Next, the system can direct you to the bot or the personal cabinet of Midjourney. To go to the bot from the cabinet, click on the avatar in the lower left corner and tap “Go to Discord.”
• Then go to any chat whose name starts with newbies: «newbies-102», «newbies-132» etc. The serial numbers are different, but it does not affect anything. The chats are on the left, on the side.

• To generate a picture from the description, enter the command “/imagine prompt” and the story’s text in the typing line, press “Enter” and wait for the result within a minute.

The text line typed “/imagine” prompt and description in English.

There are four options, and we choose the most successful (U2).

The neural network improved it and added details.

• Type and send the command “/Settings” to process your photo. A window will appear in which you must select the settings, as shown in the screenshot.

Lifehack: to get an image with a specific aspect ratio, for example, 5:4, add “–ar 5:4” (or any other percentage, the main thing is to use integers) at the end of the request.

• If the network did not suggest accepting the usage rules up to this point, write “/Imagine,” hit “Enter,” type any text, and send it. Now you should see a window with the rules, accept them, and click the green button.

• Next, you must take a photo (preferably a well-lit one) and upload it to the “ltdfoto.ru” photo hosting.

• Copy the “Direct link.”

• The finish line! Returning to the chat room Discord, type “/Imagine”.
• In the “Promt” line that appears, insert a direct link to the photo, and add a space and a short description of what we want to get (an example of the description on the screenshot – necessarily in English).

Tips for the description: If you want a photo in the style of Miyazaki cartoons, write “Studio Ghibli art style.” Write “Universe name” + “Character” if you want a character style. Then follow the same logic. Experiment – the result is limited only by your imagination.Wait for the generation of the photo for about a minute, watching the window until the load reaches 100%

There is a private processing function, but it is only available by subscription.

• The neural network generates four images at a time; you can choose the most successful one and increase its resolution by pressing buttons U1-U4, according to the number of the image. Buttons V1-V4 allow you to generate four additional variants similar to it based on one of the received images.

The neural network doesn’t just increase the resolution; it also finishes drawing small details so that the final result may differ slightly.

• If you do not like the suggested pictures, restart the generation by clicking the “refresh” icon
• Click “Open Original” and save the picture from the window that opens in your browser.

Congratulations! You get fancy avatars from AI, and you can spend the $5 you saved on your favorite coffee and dessert!