Titikey
Home实用技巧ChatGPTMidjourney User Guide: Generate Images on the Web, View Tasks, and Download the Original Image

Midjourney User Guide: Generate Images on the Web, View Tasks, and Download the Original Image

3/19/2026
ChatGPT

This Midjourney user guide focuses only on practical web-based operations: how to log in, how to start a generation task, where to view your history, and how to download the original image. Many people get stuck at “I can get into the website but don’t know where to start”—follow the steps below once and you’ll be able to use Midjourney smoothly. A few commonly used settings are also added at the end to help you use Midjourney more conveniently.

1. Logging in to Midjourney: Authorizing with Discord is the key step

After opening the official Midjourney website (midjourney.com), clicking “Sign in” will usually redirect you to the Discord authorization page. If you don’t have a Discord account, register and log in on Discord first, then return to Midjourney to continue authorization. Once authorization is complete, Midjourney will link your account information and generation history to the Discord identity used for this login.

If you find that the page is blank after logging in or that buttons don’t respond, first check whether your browser is blocking third-party cookies or pop-ups. Add midjourney.com and discord.com to the allowlist, then log in again—this usually restores normal access to Midjourney.

2. Starting a generation on the web: From entering prompts to upscaling

After entering Midjourney, go to the “Create” page. Type your prompt directly into the input box, and press Enter to start generating. For more consistent results, add details to the prompt such as the subject, style, camera, lighting, and materials to help Midjourney understand your intent. After generation, a 2×2 grid preview will appear; click a single image to open a larger viewing page.

On the single-image view page, you’ll see common buttons such as “Upscale” and “Vary.” Upscale is typically used to obtain a usable final single image, while Vary is suitable for rapid iteration within the same style, continuing to have Midjourney produce versions closer to what you need.

3. Viewing tasks and history: Find every image you’ve made

Midjourney’s generation history is generally arranged chronologically in the “Create” feed; scroll down to reach older tasks. When searching for an image, it’s recommended to use the browser’s built-in search or to locate it by recalling the time period when it was created—this is more convenient than repeatedly regenerating. If you switch Discord accounts and then log in to Midjourney again, your history will change along with the account; this is the most common reason for “missing” works.

Some accounts can also see queue/mode indicators in Midjourney (such as faster or slower generation methods), which depends on account permissions and subscription status. You don’t need to configure anything extra; just submit tasks normally and Midjourney will queue and process them according to the resources available to your current account.

4. Downloading the original image and common settings: Save finished works locally

To download an image, open the upscaled single-image page, then look for “Download” or use the browser right-click “Save as.” If you downloaded a preview image, it’s usually because you’re still on the 2×2 grid page; be sure to enter the single finished image in Midjourney before downloading—the clarity will be noticeably different. After downloading, it’s recommended to also save the corresponding prompt for reuse in Midjourney next time to replicate the same style.

For everyday use, it’s recommended that you stick to a fixed “prompt framework” (subject + style + details + aspect ratio), changing only key variables each time. This not only makes Midjourney’s outputs more consistent, but also makes it easier to compare iteration directions in your history and quickly pick the image that’s closest to the final shot.