This guide will cover how to use MidJourney to generate images from text.
What Is MidJourney?
MidJourney is an artificial intelligence (AI) program that creates images from text descriptions. It uses a machine learning algorithm trained on a lot of image data. The end result is a program that can create a wide range of art forms, from realistic to abstract styles.
Midjourney is a leading AI art creator alongside Dall-E and Stable Diffusion. It uses language and diffusion models to create remarkable and sometimes startlingly lifelike outcomes.
Currently, MidJourney is operated by a small team of around 11 people.
At the time of this writing, the only way to work with MidJourney is through an app called Discord.
What is Discord?
Discord is an instant messaging platform similar to AIM (AOL Instant Messenger) or Twitter (depending on how it’s used). Users create accounts and then can message other user or ‘bots’. A bot is usually another computer or program that can be messaged just like another human user.
How Do I Work With MidJourney?
The MidJourney program acts just like another user on the Discord platform (called a bot). To generate images, you send message commands and prompts to the MidJourney bot.
To use the MidJourney bot on Discord, you will first need to create a Discord account and join the MidJourney server.
Click Here For Help Getting Started With Discord
Once you are on the server, you can follow these steps:
- Click on the MidJourney Icon
- Go to the #newbies channel.
- Type the following command:
/imagine prompt: - Replace
prompt:with a description of the image you want to create. - Press enter.
The bot will then begin generating images based on your prompt. This process can take a few minutes, depending on the complexity of your prompt. Once the bot has finished generating images, it will send you a link to a private Discord channel where you can view and download your images.
How To Use MidJourney
MidJourney provides a command list on their website here:
https://docs.midjourney.com/docs/command-list
For the most part you will use the Imagine command. This is the command that takes a string of text as an input and uses it to generate images.
Using /imagine
The Imagine command is the most used command. It takes a string of text as an input and uses it to generate an image. The string of text is called the ‘prompt’.
Here are some tips to consider when creating a prompt:
- Be as specific as possible in your prompts. The more specific you are, the better the bot will be able to understand what you want it to create.
- Use keywords that are relevant to the style of image you want to create. For example, if you want to create a realistic image, you might want to use keywords like “photorealistic” or “realistic.”
- You can also use negative keywords to tell the bot what you don’t want to see in your image. For example, if you don’t want to see any people in your image, you might want to use the negative keyword “people.” This is typed in as ‘–no people’
- If you are not happy with the results of your first prompt, you can try generating more images by using the
/imaginecommand again.
MidJourney generates 4 images at a time. So for each prompt you submit, it will return 4 images tiled as a single image. If you want to upscale one of the images, you click on the ‘U1’, ‘U2’, ‘U3’, or ‘U4’ buttons to upscale any of the 4 images in the tile. You can also generate variations of any of the 4 images by using the ‘V1’, ‘V2’, ‘V3’ or ‘V4’ buttons.
Here is an example of a prompt to generate an image of a blue waveform.
/imagine two music waveforms on dark background, in the style of dark sky-blue and blue, 32k uhd, columns and totems, analytic art, social network analysis, glass as material, random ravelsThis generates the following image:
Using /imagine to blend images
It is possible to use an image, or a link to an image as a prompt to MidJourney. In fact, multiple image links may be used. However, you must follow the order of parameters as laid out in the parameter guide here:
https://docs.midjourney.com/docs/parameter-list
As an example, we can ask MidJourney to imagine a blend of the discord icon and midjourney icon.
First, find the urls to each icon:
- https://upload.wikimedia.org/wikipedia/commons/e/e6/Midjourney_Emblem.png
- https://logodownload.org/wp-content/uploads/2017/11/discord-logo-1-1.png
Create the prompt:
/imagine https://logodownload.org/wp-content/uploads/2017/11/discord-logo-1-1.png https://upload.wikimedia.org/wikipedia/commons/e/e6/Midjourney_Emblem.pngDiscord converts the links to shortlinks. Then MidJourney generates the following image:
Using /describe
The describe command requires an image as an input. When you type /describe into the chat, MidJourney will respond with a box to upload an image into.
/describe [drag and drop image into discord chat]
I chose to use an image from the Wikipedia article on ‘Trees’.
MidJourney generates 4 descriptions for the image. It also provides a link at the bottom to ‘Imagine All’. Clicking this button would cause MidJourney to use the Imagine command on each of the 4 image descriptions it provided.
Conclusion
In conclusion, Discord and MidJourney are two powerful tools that can be used together to create stunning and imaginative images. By following the tips in this blog post, you can learn how to use these tools to create the images you have in mind.
Here are some additional tips for using Discord and MidJourney:
- Join the MidJourney Discord server and participate in the community. There are many helpful people on the server who are willing to share tips and advice.
- Experiment with different prompts and settings to see what kind of results you get. MidJourney is a powerful tool, but it takes some practice to learn how to use it effectively.
- Don’t be afraid to share your work with the MidJourney community. The best way to learn and improve is to get feedback from others.
I hope this blog post has been helpful. If you have any questions, please feel free to leave a comment below.