Midjourney 2026 complete introductory tutorial, only 5 steps from registration to the first picture

If you are interested in AI-generated images, Midjourney is one of the most worthwhile tools currently. It does not require you to know programming or install a complex environment locally. As long as you have a Discord account and a subscription, you can generate fairly high-quality images in a few minutes. This article breaks down the process from scratch to the first picture into 5 steps. Each step clearly explains the operation details and precautions to help you get through the entire process in the shortest time.

1 What is Midjourney and why is it worth learning?

Midjourney is an image generation tool based on artificial intelligence. The user inputs a text description (that is, prompt), and the system will generate the corresponding image based on the description. Its core advantage lies in the high quality of pictures and diverse styles. Whether it is realistic photo style, illustration style or conceptual art style, it can achieve a relatively high degree of completion.

Compared with other similar tools, Midjourney has a relatively low barrier to entry. It runs on the Discord platform and does not require you to have a high-end graphics card locally, nor does it require installing Python or configuring various dependencies. This is very friendly to creators with non-technical backgrounds.

As of the writing of this article, Midjourney has gone through multiple version iterations, with significant improvements in image quality and generation speed. In the social media and design community, you can often see works generated with Midjourney being used in concept design, social media graphics, e-commerce materials, and even art resources for independent games. The application scenarios of this tool are much wider than "playing with AI drawing".

2 Step one, register a Discord account and join the Midjourney server

Everything in Midjourney is done on Discord, so the first thing is to have a Discord account. If you don't have Discord yet, open discord.com and follow the prompts on the page to register. The registration process requires email verification, and the entire process can be completed in a few minutes.

After registration is complete, you need to join Midjourney's official Discord server. Open midjourney.com in your browser. There will be an entrance to join Discord on the page. After clicking, it will jump to Discord and automatically enter the Midjourney server.

After entering the server, you will see many channels. Newbies are advised to browse the getting-started and announcements channels to learn about the latest rules and usage instructions. The community of Midjourney is very active, and people are generating pictures in the channel at any time. You can first see how others write prompts and accumulate some feelings.

3 Step 2: Subscribe to the Midjourney paid plan

In the early days, Midjourney provided a free trial, but later adjusted the policy and new users needed to subscribe to a paid plan to use it. For specific prices, please check the latest pricing page of the official website. Because Midjourney’s pricing strategy will be adjusted from time to time, specific figures are not listed in this article to avoid being out of date.

The subscription entrance is on the Account page of Midjourney's official website. After logging in to your Discord account, you can see the available packages. The differences between different packages mainly lie in the available monthly generation time (fast hours), whether privacy mode (stealth mode) is supported, and the number of concurrent generation, etc.

For users who are just getting started, choosing the most basic package is enough. Go through the process first to confirm that you really have the need for continued use, and then consider upgrading. After the subscription is successful, you can officially use the generation function of Midjourney in Discord.

4 The third step is to use the /imagine command to generate the first picture.

After the subscription is completed, return to Discord's Midjourney server, find any newbies channel (novice channel), enter /imagine in the message input box and press Enter. Discord will pop up a prompt input box, where you can enter the image description you want and press Enter to send.

For example, you can enter a description like a cozy coffee shop in Tokyo, morning light, watercolor style. Midjourney will generate four candidate images in about a minute and display them to you in a 2x2 grid.

There are a few points to note here. prompt currently mainly supports English input. Although it can also recognize part of Chinese, the accuracy of understanding English is obviously higher. If your English is not very good, you can first use a translation tool to translate your Chinese description into English, and then paste it into the prompt. In addition, the prompt does not need to be written like a complete English sentence. The combination of keywords and phrases is more effective.

5 Step 4: Understand the prompt parameter and control the generation effect

Midjourney's prompt is not just a text description, you can also add various parameters after the description to precisely control the generated results. Mastering these parameters is a key step from "playing around" to "producing usable works".

The most commonly used parameter is --ar, which is used to set the aspect ratio of the image. By default, a 1:1 square image is generated. If you need a horizontal image, you can write --ar 16:9, a vertical image, --ar 9:16, and a mobile wallpaper, --ar 2:3. The aspect ratio has a great influence on the composition. It is recommended to think clearly about the purpose of the image before deciding on the proportion before each generation.

The --q parameter controls build quality. The default value is 1, you can set it to 0.5 to speed up the generation (slightly lower quality but faster, suitable for quick test prompts), or you can set it to 2 to get higher quality results (which will consume more generation time).

The --style parameter can switch between different style tendencies. Different versions of Midjourney provide different style options. It is recommended to consult the official documentation for specific available values.

The --no parameter is used to exclude elements you don't want. For example, if you generate a landscape but don't want people to appear, you can add --no people. This parameter is very practical in actual use and can help you avoid many unwanted picture elements.

A complete prompt example: a futuristic city skyline at sunset, cyberpunk atmosphere, neon lights --ar 16:9 --q 2.

6 Step 5: Enlarge, morph and save your image

Two rows of buttons appear below the 2x2 grid map generated by Midjourney. U1 to U4 are Upscale (enlargement), corresponding to the four pictures in the grid. After clicking the U button, Midjourney will enlarge the picture you selected with high resolution and generate a larger and clearer version.

V1 to V4 are Variations. After clicking, Midjourney will generate four new pictures with similar styles but different details based on the picture you selected. When you feel that the direction of a certain picture is correct but the details are not satisfactory enough, you can use the variant function to quickly iterate.

There is also a refresh button. When clicked, four new pictures will be regenerated with the same prompt. If you are not satisfied with all four pictures, you can directly refresh and start again.

The method of saving pictures is very simple. Click on the picture you want to save in Discord, the picture will be enlarged, then right-click and select "Save Image". If you need to manage and download all generated images, you can log in to your personal homepage at midjourney.com, where all your generated records will be saved in chronological order.

7 Practical Tips for Writing Better Prompts

The quality of prompt directly determines the rendering effect. The following tips can help you improve quickly.

First, describe the subject first and then the environment and style. For example, first say "an orange cat", then "sitting on the windowsill", and finally "soft afternoon sunlight, film photography style". This structure allows Midjourney to understand your intentions more accurately.

Second, make good use of specific style keywords. For example, oil painting, watercolor, 3D render, anime style, cinematic lighting, etc. The more specific style words are, the more controllable the generated results are.

Third, refer to other people’s excellent tips. There are a large number of high-quality prompt cases in the Midjourney community and various prompt sharing websites. In the early stage, you will watch and imitate more, and gradually form your own prompt style.

Fourth, iterate rather than do it all in one step. Very few people get perfect results with just one prompt. The normal workflow is: write the initial prompt, look at the results, adjust keywords, regenerate, select variations, and get a satisfactory work after repeated rounds.

8 common mistakes newbies make

The first common mistake is to write prompt too generic. For example, just write "a beautiful girl". This description lacks specific information. Midjourney doesn't know what style, scene, or light you want, and the generated results will be very random.

The second mistake is to cram too much into the prompt at once. Putting more than a dozen elements into one prompt often results in a messy picture, with no element performing well. It is recommended to focus on 2-3 core elements at a time.

The third mistake is ignoring the aspect ratio. The default 1:1 square is not suitable for all uses. If your image is used as a mobile wallpaper, web banner, or social media cover, failure to set the appropriate aspect ratio will result in cropping and loss of the image later.

The fourth mistake is never using the variations feature. Many novices rewrite the prompt when they are not satisfied with the first set of pictures. In fact, a more efficient way is to fine-tune the point variations that are close to satisfactory on the existing basis.

The fifth mistake is not to read the community and official documentation. Midjourney is updated frequently, and new versions often add new parameters and features. If you don't read the documentation, you may have been using outdated writing methods and missed better generation methods.

For readers based in mainland China, the Discord registration plus international card subscription plus VPN combo is often the part that blocks first-time users. A lighter on-ramp on iPhone is an app called Lingtu (灵图), full name "灵图-AI画图设计" on the China region App Store. It is downloadable directly without a VPN, takes prompts in Chinese, and bundles several overseas-style engines under one interface, including a Midjourney-style ambient engine, a Flux-style photoreal engine, and a Nano Banana-style fast engine. It is a practical place to get a feel for prompt structure and parameter behavior before committing to a full Discord plus Midjourney subscription setup.

FAQ

Can images generated by Midjourney be used commercially?

According to official public information, images generated by paid subscribers can be used for commercial purposes, but it is recommended to check the Terms of Service page of Midjourney's official website for specific usage rights and restrictions, as the terms may be updated over time. Images generated during the free trial may have limited commercial permissions.

Can I use Midjourney if I don’t know English?

It can be used, but the experience will be compromised. Midjourney has the most accurate understanding of English prompt. It is recommended to use a translation tool to translate the Chinese description into English before entering it. There are actually only a few dozen common style words and scene words in daily use. You can remember them after using them a few times and do not require a high level of English.

Which one should I choose, Midjourney or Stable Diffusion?

The two positions are different. Midjourney is easy to get started, does not require local graphics cards and environment configurations, and has stable image quality, making it suitable for designers, content creators and non-technical users. Stable Diffusion is an open source solution that is highly customizable and supports local deployment and model fine-tuning. It is suitable for users who have a technical foundation and require in-depth control. If you are a novice, it is recommended to start with Midjourney first, and then consider Stable Diffusion as needed after becoming familiar with the basic logic of AI graph generation.

How long does it take to generate a picture?

Under normal circumstances, it takes about a minute to generate a set of four images using fast mode. Using relax mode (supported by some plans) may take a few minutes, depending on the queue situation at the time. Upscale usually takes less than a minute to enlarge an image. Overall it's very fast and won't keep you waiting too long.

Can you use Midjourney to generate photos of specific real people?

Midjourney's community rules prohibit the generation of unauthorized portraits of real people, including public figures and ordinary people. If your prompt contains real names, the system may refuse to generate or produce ambiguous results. This is a restriction due to privacy and ethical considerations. If you need pictures of characters, it is recommended to use fictional descriptions and let AI generate original virtual characters.