Candy AI Image Generator: My Unfiltered Thoughts

Candy AI’s NSFW AI image generator occupies an intriguing niche somewhere between tool and sandbox. It’s not like you are making random, arbitrary, adult AI pictures; it’s more like you are designing a character, setting up scenarios, and playing around with details until you like what you see.

From my brief exploration (I know, I spent longer on Candy AI than was reasonable), the experience is pretty simple: choose a character, write a prompt, pick some styles and details, press generate, and then either walk away happy or become slightly addicted.

As the name suggests, Candy AI is geared toward personalized NSFW content as opposed to generic generated AI images.

How to use Candy AI: Step-by-step guide

This image generator has a 3-step workflow:

• Pick a character
• Enter the prompt, click generate.
• Create an account

The user interface is quite user friendly. It does not present everything at once so that is a relief. Access, then pick a character, then create. Very straightforward.

Below is a detailed guide of the screenshots and all the interface elements.

Step 1: Pick a Character

The next step in the interface is the character selection screen, giving you the opportunity to choose a base visual to begin the generation process. This is a very crucial step, since it forms the subject of the image you want to generate.

On the screen, you will see the following:

Generate Image Choose character

This is the main heading for the screen you currently are in, letting you know you’re in the generation process but need to first pick the character.

On this screen there are several things worth noting:

1. Character Library Tab Categories

On the upper left, there are two tab headers which indicate the types of characters available for selection:

• Anime
• Guys

The first tab (Anime) is highlighted, indicating that this is the currently selected one.

This means you are able to sort the library by different categories, likely by character style or gender.

Which means:

• If you want to see anime characters, choose Anime.
• If you want to see masculine characters, choose Guys.

These are the only two category types displayed, although the rest might fall outside the current view.

2. Character Library

The remaining main screen features a character selection grid, giving you a quick preview of all the characters available to you. This can be useful for seeing different character options before you choose which character you want to use for the generation.

Here’s what the grid is showing on this view:

• Nari
• Yama
• Rosalie

These are the fully visible choices currently displayed in the center, although there are also partially visible characters on the sides of the screen as well as on the bottom, indicating there are definitely characters available to view that are not currently visible.

Each card also has the following features:

• An image portrait of that character
• The character’s name
• And the rest of the card is a clickable selection.

3. Selected Character

The selection is highlighted on the grid. Yama is highlighted in this screen shot.

This is great, it indicates:

• The character is selected for you.
• The system is ready for the next step.

There’s no guesswork, no clicking around three times to ensure you actually clicked something.

4. Choose Character Label

The heading for this step helps you to understand what this part of the system is for.

Which means:

• You’ve selected the right tool.
• You have now selected this part of the tool.

The arrow in the image is pointing directly towards this area, as a way to guide you to this.

5. Select Button

Lastly, there is the main Select button towards the bottom of the screen:

Select

There is also a large arrow in the lower left corner pointing to it, highlighting that this is the next thing you should select to take the next step in the process.

How to complete Step 2

• Look for your chosen character card in the library
• Select the Anime or Guys filter at the top if you wish
• Click your desired character to select it
• Check that the card highlights
• Press the Select button

Tips for picking a character

The character is not just decoration. It’s the visual anchor of the image you generate later.

Step 2: Write the Prompt and Generate the Image

The third screen shows the generation editor, in which the user describes the image desired and indicates how many images to produce. This is where the fun happens, or where you learn that your prompt wasn’t clear enough. (Likely both.)

What you see on this screen

The interface contains several key sections:

Features visible on Step 3

1. Character selection preview

On the left you’ll see the character that has been selected. For instance, the screenshot here features:

Yama Gito as the selected character

The feature is useful because it will remind you what character is selected when you’re generating.

Note that the selected character card also has a small icon located in the top right of the card. Though we don’t see a clear label for what this does, a safe assumption is that this will be an interactive control that relates to the character. It might allow you to view or change things related to that character.

2. Generate Image

Also, at the top of the page, you’ll see the header:

Generate Image

This confirms you’ve reached the generation step.

3. Prompt input field

The right-hand side of the character preview shows a text box which has a default text prompt:

Sitting on a leather sofa, wearing a fur jacket, wearing lace underwear, gazing seductively at the viewer.

This confirms that the application is controlled by a written prompt.

With the prompt box you are able to describe:

• The pose of the character
• The setting or background
• What the character is wearing
• The facial expression or mood of the character
• The general atmosphere of the photo or drawing
• The point of view or angle

How to use the prompt box

A good prompt usually includes a few basic ingredients:

A clean prompt structure may look like this:

[pose/action] + [setting] + [outfit] + [expression/mood] + [camera direction]

Example:

sitting on a leather sofa, wearing a fur jacket and lace underwear, gazing seductively at the viewer

That is a straightforward and clean setup.

4. Suggestion tabs

Just below the prompt is a series of tabs:

• Suggestions
• Outfit
• Action
• Pose
• Accessories
• Scene

These are prompt helpers, which help the user find ideas or add relevant prompt details.

What each suggestion tab likely helps with

This feature is great for new users. Often, you have an idea of what you want, but you can’t find the words to express it. These tabs help to fill in any holes.

5. Suggestion chips

Underneath the tabs are multiple clickable chips, such as:

• bikini
• skirt
• lingerie
• crop top
• leather
• mini-skirt
• satin r… (cut off)

It looks like these buttons serve as a quick way to help you expand or enhance the text. How it could be used:

• To quickly add clothes
• To get an idea of what to type
• To make writing the prompt faster
• To not be faced with a blank text box

6. Image counter

Underneath the suggestion area is a label that says Number of images. The choices you see are:

• 1
• 4
• 16
• 32

The first number is selected. This is to select how many images the AI will generate together.

What each image count is good for

Some of them even have smaller symbols on top, most likely indicating a credit, token, or premium cost. That generally means that larger quantities will cost more credits. Therefore, when using this for the first time, selecting 1 is the best choice.

7. Generate Image

At the bottom, there is a large gradient button that says:

Generate Image

This is the action button that you’ll need to press at the very end. After selecting a character, writing a prompt, and choosing the quantity of images you want, clicking this button will initiate the creation of the image(s). The screenshot points right at it, so this is obviously the final step.

Step 3: Create an Account

The website asks you to register or log in before being able to generate images.

The page can be broken down as follows:

• Left side: The site shows a woman and the website’s branding elements.
• Right side: The signup form and the login buttons.

The site seems to have kept this section extremely simple. As the arrows in the image above indicate, there are only two steps required in this section.

The things we see in the first section:

1. The Header: Create Account.

You see the words Create Account at the top of the section. This tells us we are at the signup section.

2. The Email field.

The first input is simply labeled Email and asks you to provide your email address. The goal is for you to create an account and for the site to be able to keep track of your creations and settings.

The red circle in the top part of the image highlights this item, indicating it as the first step.

3. The Password field.

The second box asks you to provide a password.

You find some text below the field saying:

Minimum 6 characters

That means your password must contain at least six characters.

As the arrow in the bottom part of the image indicates, the field is the second step.

4. The Create Free Account button.

The button at the bottom part of the section is the main Create Free Account button. Once you’ve provided your email and password, you click the button to complete the signup.

5. Sign up with.

There’s another line of text followed by a horizontal bar below the Create Free Account button:

or continue with

Below that text is a set of three social sign-in buttons:

• Google
• Discord
• X

You can use these if you’d prefer not to manually enter your credentials.

A good option, if I do say so myself. Some folks might not want to keep adding to their 47 passwords this year.

6. Terms of Service notice.

At the bottom of the signup form is also some text telling you that your agreement to their terms of service is binding once you’ve completed the signup. It’s the same disclaimer used in most signups everywhere.

7. Already have an account? Sign in.

Finally, at the bottom of the page, you find yet another option, this one is to Sign In, in case you already have an account.

8. The left side preview and the message regarding your choices.

On the left side, you see a big preview of the woman with a text saying:

Your choices will be saved.

That’s probably referring to the fact that the platform will save your characters, settings, and choices you’ll make (perhaps even your image history) once you are signed in.

The candy.ai logo is also visible in the bottom left.

Steps in order to complete the first section:

Option 1: Create an Account

• Enter your email address
• Enter a password that is at least six characters long
• Select the Create Free Account button

Option 2: Sign up or log in with social media accounts.

• Select one of the three login options: Google, Discord, or X
• Go through the prompts
• Log into the platform

Option 3: Sign in.

If you already have an account on candy.ai, you can choose this option:

• Click the Sign in button
• Log in with your current credentials

Step 3 summary

A Complete Start-to-Finish Guide

1. Step 1: Choose a character Go to character selection screen Use Anime or Guys to filter by category Click a character card Confirm it is highlighted Click Select
2. Step 2: Write description and generate images Check out the character on the left Write your prompt in the text box (e.g. sitting on a white sofa, looking at the viewer) Utilize Suggestions, Outfit, Action, Pose, Accessories, and Scene Click on prompt chips to add ideas Select the quantity of images Click Generate Image
3. Step 3: Create your account Type in your email address Type in your password Press Create Free Account Or press Google, Discord, or X Or press Sign in if you already have an account

Tips for Getting Better Results

Start with just one image

Pick 1 image to start.

This allows you to refine your prompt first before burning more credits for bigger quantities you might not even like.

Layer in details when writing prompts

Don’t write everything at once in the first try.

Try starting with:

sitting on a sofa, looking at the viewer

Then build up the prompt and try again:

sitting on a leather sofa, wearing a crop top, looking at the viewer in warm lighting

You’ll see that this approach works a lot better than stuffing half of a book into a single prompt field.

Leverage the tabs if you’re at a loss for words

If you can’t think of what to write down, try using:

• Outfit
• Pose
• Scene

These are tab options available for you for a reason. Let them help you out!

Keep using the same character if you want consistent faces and styling

If you want multiple images of the same person, use the same selected character.

This should help ensure consistency with:

• facial features
• art style
• overall identity

Once the prompts are working, increase the batch size

Once you’ve started receiving results with 1 image, increase the number to either:

• 4 for variation
• 16 if you want even more results

It’s an extremely risky move to go straight to 32 without prior testing. Some may say it’s a little too risky.

Wrapping Up

This image generator uses a very logical process:

• select the person you want to create
• enter what you would like to generate
• log into the service

This order keeps the experience simple and easy for a new user.

Some key features that stand out visually are:

• the easy to follow registration
• the grid of characters
• the prompt options
• the batch selector
• the big create button

The overall design is user friendly yet powerful for someone with experience. This can be turned into a concise how to guide, a blog post, or an easy to digest product guide if you would like.