Midjourney AI - Tutorial for Beginners in 12 MINS! [ FULL GUIDE 2024 ]

Dear all, welcome to this video on Midjourney, an outstanding AI content generator able to craft images from given prompts in a Discord environment! This is available under a subscription plan with no more free versions available. In just 12 minutes we are going to see how to run it on a private server and how to use it from scratch! To start using Midjourney you need to download and install Discord in your computer and create an account to log in. At this point you can open the Midjourney page

and click on Join the Beta to create a Midjourney account that will be automatically linked to your Discord account. This will show the Midjourney server and several newbie rooms you can use to make experiments while testing this bot publicly! You can also run Midjourney on private servers for an additional layer of privacy. Go to Add a Server on the left and then to Create My Own. Locate the Midjourney bot, click on it and go to Add to Server to add it to your own private server and be ready to

generate your first images through dedicated prompt commands. Pay attention that all images generated with Midjourney allow public use unless you have any plan including the Stealth Mode. Check our video description for more information! Every time you generate any image with Midjourney you consume your GPU monthly time from your subscription plan. The amount consumed depends on the chosen performance between Relax, Fast and Turbo Mode that change the speed of the generation without impacting i

ts final result. Check the Settings chapter to see how to choose the right Mode for you. You may go for a Standard subscription plan in order to have unlimited GPU time in Relax mode and go on using Midjourney without limits! Now let's see how to generate outstanding images with Midjourney. This bot reads textual prompts you give as instructions made of a main command, a proper description and eventually additional parameters to configure and customize the result. For example you can use the /im

agine command and describe a fruit basket to generate four versions of images showing this subject. You can click on any to get a wider preview and choose any button from U1 to U4 to upscale the resolution of any image, counted clockwise from the top left image. This also saves the chosen image as the final one. Let's generate a smiling girl image. You may use the V1-V4 Variation buttons in clockwise order to remix the chosen image and relaunch the algorithm to improve the description and apply

modifications, for example making the image more colorful. To save the image obtained go to Web, to the three dots in the top right corner and hit Download. You will get a PNG image with the default size of 1024x1024 pixels! Now let's see how to write a good prompt. Specify the main subject, the environment and the background and also the wanted style, such as a photography, an illustration or a painting. You can also define more details like vibe, kind of lights or colors to use. For example yo

u define a girl as subject, over hills in a colorful style. Use V to generate the image again by changing its style to a realistic black and white photo. You can also generate customized logos, just pay attention that Midjourney is not able to handle text precisely and it always saves your images in PNG format, with no vector formats available. Once you selected an image as final from variations you get further options at the bottom. Select Vary(Strong) to edit the image through additional promp

t instructions, maintaining the original mood and background. Choose Vary(Subtle) to perform minor modifications and select Vary(Region) to edit the image only on specific regions. In this case enable the Rectangle tool at the bottom click and drag to define the interested area, then specify the prompt at the bottom. Your instructions shall match the subject and the background within the area to get the best results. You can also use the Upscale option to increase the size of the image; Zoom Out

to zoom out from the image and click on the Pan arrows to extend the image through any specific direction chosen. In Midjourney you can also extract the prompt matching a given image, operation called reverse prompting. Enter the /describe command, click to browse for the image and get four versions of the possible prompt. This way you can generate a new image or go to Imagine all to generate all of them! You can also use the /shorten command to optimize your prompt by removing useless words. M

ake sure to double-check it or the subject of your generation may change. If you want to specify multiple subjects on an image you may use the multiprompting with two colons. For example if we want bedclothes we obtain this kind of image, but if you type bed:: and then clothes, you get images featuring bed and/or clothes as subjects. You can also add the number 2 after the colons to rise the importance of the first subject over the second one. Midjourney supports permutation, such as the generat

ion of arrays of images following a list. For example we can generate three humanoids specifying the list under brackets and separated by commas, obtaining four versions per each element of the array. This feature works in Turbo Mode only. Midjourney can use real images as reference to generate others. If you have one online you can copy its link next to the /imagine command with your own options and launch the prompt. You can also import an offline image by going to the plus button and then to

Upload a File. At this point right-click on it and go to Copy Link to use it with the /imagine command! With Midjourney you can also merge images together! Just type the /blend command and upload up to five images to mix together in style and content! Now let's see how to use the command parameters. These are simple keywords or letters after two hyphens added at the end of the prompt to change the effect of the main command. The version (v) parameter allows you to select a different AI algorithm

version, adding the 'style raw' parameter to generate images that better match your prompt. Use the stylize (s) parameter to balance the creativity of the AI algorithm with default value of 100 and maximum value of 1000. A lower level brings images closer to the prompt, while a higher level gives more freedom to the algorithm. With the chaos (c) parameter you can increase the randomness level of your generations up to a maximum value of 100, even over the prompt description. For example if you

take a lions image a higher chaos parameter may vary the main subject, up to showing fantasy features! Use the quality (q) parameter to set the generation quality between 0 and 1. A higher quality takes more time to generate the image but with more details and less blurriness. With the no parameter you can add negative prompts, in order to define what you do not want on the generation. Use the tile parameter to add more elements within the same image, in order to make patterns or textures. For e

xample you can generate any strains and verify how this can be replicated with any Seamless Pattern Checker website you can find online. Use the weird parameter to add strange and eccentric elements on your images and get unexpected results, with values from 0 to 3000. With respect to chaos, this parameter always follows the prompt instructions. The aspect (or ar) parameter sets the aspect ratio of the generated image through width, colon and height as ratio. Whereas use the seed parameter to co

ntrol the randomness of the generation by providing a seed number: generations with the same seed number provide very similar results. In case you want to retrieve the seed number from a previous generation just send this one to yourself with the envelope, copy the seed number and use it to get images with the same style. If you want to learn more about these parameters take a look at the official page link you can find in the video description! There are other useful commands you can use in Mid

journey. With the /info command you can check your account and subscription status. Launch /settings to configure your generation defaults, such as any suffix to apply to your prompts in this case the stylize parameter value. You can also set the algorithm model to use, enable Public Mode to make your images free to use and the Variation level when you remix images with the V buttons. You can also enable any Mode between Relax, Fast and Turbo for additional GPU speed and also choose a different

AI model such as the Niji model to realize images following any anime, manga or cartoon style with its own set of options. If you need more help and information you can visit the documentation site of Midjourney, showing tutorials and step-by-step instructions. For even more hints and tips you can also visit the Promptomania portal to get support from a huge community working on AI image generation! Thank you very much for watching this video! Make sure to visit our channel and our website to le

arn more about AI tools and enhance your work through our complete and free guides!