What is niji・journey?

This is a two-dimensional AI painting model jointly launched by MidJourney and Waifu.

How to qualify?

Apply through https://nijijourney.com/zh/getting-started/ , you can get 25 free trials.

How to view the generated pictures?

Log in with Discord at https://www.midjourney.com/app/ to view past generated images and related information.

What is the price of niji・journey?

$10/month is limited to 200 copies, $30/month is unlimited, but more than 1,000 copies will be speed-limited, and an additional $20 can obtain commercial copyright and privacy mode.

What is Privacy Mode?

Interacting with the bot alone and private mode is not truly private, others can still view and retrieve it. True privacy is only available at an additional cost.

How do I fill in the descriptors (PROMPTS)?

Chinese is supported, and the system will use the deepl translator to translate it into English and then pass it to the algorithm. Minor vocabulary may be mistranslated. It is recommended to try to translate first, and then modify the parts that may be wrong in English.

How many pictures can be generated at the same time each time?

Each user can have up to three image generation threads.

What is the most basic method of operation?

Enter a slash/imagine in the dialog box with a bot robot or directly click the pop-up command after entering a slash.

Getting Started

Use /imagine

In any #imagegen-related channel, enter the command /imagine with the prompt text, and the spawnbot will start the build for you. The AI replies with intermediate results from an ongoing generation that will slowly update to four full images over a few minutes.

Adjust initial results

The interface after generating the picture is as shown in the figure below. Each picture is by default composed of four pieces of 512x512 spliced to a size of 1024x1024, and the picture number is as shown in the figure. U stands for Upscale, which can enlarge and optimize the corresponding picture. But this optimization does not simply enlarge 512 to 1024, but is equivalent to regenerating the picture. Therefore, the enlarged content may be different from the small details in the four original images, and may be better or worse. The --uplight parameter can be used to make fewer changes during uplighting. V stands for Variations, which will generate four new images based on the selected image, and try to maintain the overall style and composition without changing the resolution. The modified content is more than U. Run again means that no matter which operation is performed in the current task, the system will re-execute it. It can produce very different results, especially when re-running the /imagine command. After clicking U, there are the following buttons below the generated single image: Upscale to Max further increases the resolution of the enlarged image (it may not be available in niji, it may only be available in mid). Make Variations is equivalent to the V function, using the Upscaled image as the basis to generate four new images, trying to maintain the overall style and composition without changing the resolution. Light Upscale Redo redoes the upscaling process in a different way, trying not to add too much detail to the image. If you want to use this method as the default magnification method, please add the --uplight parameter at the end of the /imagine command. Beta Upscale Redo uses a more efficient way to do upscale repaints, creating higher resolution images from the start. If you want to use this method as the default magnification method, please add the --upbeta parameter at the end of the /imagine command. Note that this feature is experimental and may be removed or undergo major changes in the future. The function of sharing and expressing emotions is used to share and express attitudes and feelings towards the generated pictures.