💕How to Use Nai on Discord
Join our server and generate amazing art with the imgnAI community!
Last updated
Join our server and generate amazing art with the imgnAI community!
Last updated
Join the imgnAI server (https://discord.gg/imgnai) and head to the AI Generations section.
You can use the #sfw channel for Safe for Work generations and the #nsfw channel for Not Safe for Work content.
Type / and the model menu will appear. Click to select a model, then type in your prompt (a description of your image). We recommend starting with /gen. This command calls GenXL, our most advanced model to-date.
Your image will be generated and appear in the channel. Each prompt generates 4 images.
Once you've entered your prompt and the initial image grid has been generated, you can use the buttons to select an image, create variations, or transform with remix/chaos.
The 'S' buttons allow you to select an individual image. From there, you can create further variations, transform with remix/chaos, or upscale the resolution.
You can use the Variations button to create variations of an image. For example, by clicking 'V1' below, we created 4 more variations of image 1 from the grid.
You can create variations from the grid, or you can select an image and then create variations using the same model.
The Remix button regenerates your prompt with random prompt modifiers. The added tags are shown in the results so you can use any you like in future generations.
You can use remix on a selected image, or use it to transform your original grid.
Similar to Remix, the chaos button regenerates your prompt with added prompt modifiers. However, the added tags are more random and meant to more drastically transform the original image.
You can use chaos on a selected image, or use it to transform your original grid.
Select and image and use the re-imagine button to re-run this image through the four models used by the /imagine command.
Advanced tags can be used to further transform your creations with Nai. Please note that some of these tags/commands are currently only for Beta Users.
Advanced tags available:
--remix : Add some randomly generated prompts to mix up your art style!
--chaos : Add a lot of randomly generated prompts to radically alter your art style!
--private : Prevent other users from seeing your input image as a thumbnail.
--lock : Lock your generation, preventing other users from generating variations of it.
Note: You can also DM Nai to generate and receive images privately, without them being posted in a channel!
--wide : Generates your image in wide dimensions (1.5:1)
--tall : Generates your image in tall dimensions (1:1.5)
--hq : Enable HQ (756x756) image canvas (BETA)
--hd : Enable HD (1024x1024) image canvas (BETA - 1:1 dimensions only)
Note: --hq can be mixed with either --wide, or --tall!
--raw : Removes Nai's Prompt Engineering (can result in low quality images - recommended for advanced users only)
--seed=x : Sets a seed x for your image (typically used for iterative work on the same generation), eg: --seed=1234
--cfg=x : Manually define a scale for how strictly Nai should follow your prompts.
eg: --cfg=16 would specify the strictest adherence, sacrificing overall quality.
--uc=[x, y, z..] : Defines unwanted tags in your image - for example, --uc=[sun, moon, stars] would explicitely ask Nai to avoid including the sun, moon, or stars in your image
All advanced tags begin with a double-hyphen, and should be added to your generation prompt as normal
For example: to use the --wide and --hq tags on the prompt pink hair catgirl, you'd simply update your prompt to pink hair catgirl --wide --hq
Commas and placement do not affect advanced tags, though we recommend placing advanced tags at the end of prompts
