My Experience with Candy AI Video Generator

The Candy AI video generator, which feels like a natural extension to those images: you have a character, the setting is established, the vibe is there, and then you go… “what if it was a video?”

It is really a similar concept to Candy AI’s AI image generator: a character, prompts, and an aesthetic vibe, but with the added element of movement, and not as a full-length scene (again, don’t get me wrong), but just enough movement to make it seem real.

I cannot say the transition from image to video, even for a few seconds, has gone without notice!

A comprehensive guide of how to use Candy AI Video Generator

As seen from the screenshots provided, this video generator works basically the same way as the image one. I noticed 3 stages on the screen:

1. Pick a Character / Face
2. Enter generation settings and click Generate
3. Register / Sign Up

Quick Overview of the Workflow

Step 1: Pick A Character

This is the second screenshot, and here you select your character that will be the base of the video generator: The second screenshot, you pick your character that will be the base of the video generator: Now it might not sound like you are choosing more than just a thumbnail in your case, but this will be what the video generator will be using for your thumbnail. On this screen it says: Generate Image Choose Character Even though this says Generate Image, I told you it is also being used for the video, so it shows that it is using the same selection process for a character for images.

1. Tabs There are tabs like Anime and Guys at the top left of the page. Anime is selected in the screenshot. This means you have your selection of characters, and this is where it is divided and there are likely more tabs, but you can just change from Anime to guys.

2. Characters This screen has multiple character cards. In this photo there are Genesi, Nari, Yama, and Rosalie. There are some that are slightly visible from the outside of the photo that you cannot see the names or image for, and likely there are more below. Each card will show you the image of the character and their name, with the option to select.

3. Your selection In this screenshot you will notice that Yama has a pink border. You now know what character you are selecting and will use in the next step. It lets you know that it clicked, which is very nice to avoid a little annoyance.

4. Choose Character This shows the user Choose character, with the arrow pointing at the top, to help them find what to click next.

5. Select This is also where you will click on Select with the white arrow pointing at the button. Select is a large gradient button at the bottom.

Here is a summary of what to do in Step 1: Browse and Click Characters You select your character from the options above using Anime or Guys. Then Click your desired character. Make sure you have the correct character by the colored border around it. Then Click Select.

Tips for choosing the right character

will shape the overall aesthetic of your end result.

Step 2: Prepare for Video Generation

You are seeing the main creation screen in screenshot 2. This is where you describe the scene and select generation options, before finally clicking the generate button.

As with the last step, even though the text says Generate Image, given you are asking for a video generator, the prompt interface is similar.

On this screen, there are different sections:

Generation Screen Features

1. Selected Character Preview On the left is a preview card with the character selected:

Yama Gito

This is a reminder for the user of who they’re generating with before pressing a button.

There is also a small icon on the top right of the character preview card. The screenshot doesn’t indicate what this specific icon does so I’m going to simply assume this is clickable character related feature such as changing or editing the currently selected character.

2. Prompt Box

On the right is a large box for text input of a prompt:

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

So, this prompt tells us that the AI will need input describing a character like this.

The text box appears to take prompt detail like the following:

• setting
• pose
• outfit
• mood
• expression
• camera direction

These are the kinds of things likely to influence the scene and style when generating a video.

The basics of a good prompt

Usually, a good prompt will consist of a few key components:

A well-organized prompt can often resemble:

[action/pose] + [place/location] + [clothing attire] + [expression/mood] + [camera angle/position]

Example:

sitting on a leather sofa, wearing a fur jacket and lace underwear, looking at the camera

For a video prompt, it’s usually best to maintain simplicity with visual description, without going into excessive details.

3. Suggestions

Below the prompt box there is a series of tabs:

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

They look like prompt suggestions.

These are probably used to browse prompt ideas or add prompts to your prompt box quickly.

The purpose of each suggestion tab

This is helpful because sometimes you know what you want visually, but your brain just hands you a blank screen and vibes.

4. Prompt chips

Below the row of tabs, there are several clickable keyword chips you can see are:

• Bikini
• Skirt
• Lingerie
• Crop top
• Leather
• Mini skirt
• Satin l… (partially cut off)

These chips are designed to:

• Accelerate the writing of prompts
• Provide inspiration for the user
• Assist in the fine-tuning of the look of the clothes

They will be especially helpful to a novice user, as one can use these chips to avoid having to create every single prompt from scratch.

5. Number of images selector

Beneath these helpful prompts you will also notice a dropdown list called:

Number of images

The following options can be seen:

• 1
• 4
• 16
• 32

At this moment the first option is the one in use.

This is the one part of the screenshot which I have to be careful and honest about in your case since even if you call this video generator, the visible label is Number of images, which leads me to speculate that the screen you see here is a shared interface and the platform is actually using image-style generation parameters for the video generation as well.

So, based on a single screenshot it is difficult to say much more than the following:

• The user can choose the number of images to be generated
• A higher number is likely to increase the cost in terms of processing power
• At the moment the user has chosen the first option

Some of the higher-count options show small icons, likely indicating:

• premium cost
• token usage
• plan restriction

6. Generate Image button

Lastly, there is one button at the bottom which is:

Generate Image

This will be the last button the user will need to interact with in these screenshots, and it is very likely the one to start the generation process. The arrow in the screenshot points straight towards the button.

How to complete step 2:

• Confirm the chosen character on the left
• Tap into the prompt box
• Write the description of your scene
• Tap into Suggestions, Outfit, Action, Pose, Accessories, and Scene in order to better refine the prompt
• Select some keyword chips
• Choose the number of images
• Click Generate Image to initiate the process

Step 3: Sign up for an account

As depicted in the image below, before a user can access and use the generator they are first asked to register or login to their account.

This screen is divided into two parts:

Here are the details you can see on the account screen:

Create Account heading: At the very top of the form is the title “Create Account.” This informs the user that this is the page to register a new account.

E-mail field: The first input box is labeled E-mail. This is where the user types of their email address.

Purpose: This establishes the user’s login identity and ties any saved selections or future generations to their account.

Password field: Underneath the E-mail box is the Password field, followed by a note stating, “Minimum 6 characters.” This is telling you that the password must contain at least six characters.

• There’s a helpful eye icon to the right of this field. Typically, this allows the user to reveal the password to verify their typing. That’s very useful given how difficult it can be to enter passwords on either mobile or desktop.

Create Free Account button: A large pink button labeled “Create Free Account” appears below the input fields. This is the main button to proceed with registration after completing the fields above.

Social login options: The next section contains a divider with the words “or continue with” followed by three sign-in choices:

• Google
• Discord
• X
• These buttons provide a quicker sign-in alternative for those who don’t want to register a new account.

1. Terms of Service notice: Just beneath the social sign-in buttons is a mention of the Terms of Service, letting the user know that by signing up, they must agree to the Terms of Service of the site.
2. Sign in link: At the bottom of the page, you see the message “Already have an account? Sign in.” This serves existing users to log in to their previously made account.
3. Left-side preview image: On the left side is a large image showing a woman on a city street. It includes the text “Your choices will be saved.” It suggests that the platform saves selections and preferences, as well as prompt history, account activity, and other information. The site’s branding, candy.ai, is visible below this image.

Step 3: How to create a new account

• Option A: Manual Account Registration: Click the “E-mail” field and type in your email. Click the “Password” field, type in a password (you must use at least 6 characters), then click the pink “Create Free Account” button.
• Option B: Social Login: Click “Google,” “Discord,” or “X” to log in with your social accounts and follow any additional login prompts, then come back to this page.
• Option C: Existing Account: Click “Sign in” and login using your previously created login info, then return here.

Step 3 summary

Step 4: Generate and Review the Final Video

This is the fourth and final step in the process and the one that is not shown in the screenshot. I would like to make this explicit now.

All we know for sure is that after clicking the main generate button, the system begins processing, and once this is done, the user should check whether they like the result and save it or maybe click on retry.

Here is what I think will happen when the user is in this stage:

Based on the visible user experience, this stage will include:

Since this particular screen won’t be displayed, I suggest referring to it as the “review and output” phase rather than labeling it a final, confirmed interface.

Complete End-to-End Process

1. Step 1: Register an account Type in your email Set a password Tap Create Free Account Or pick from Google, Discord, or X Or choose Sign in if you’re already registered
2. Step 2: Pick a persona Go to the character picker Filter by Anime or Guys Pick a character Look for the pink highlight around your choice Press Select
3. Step 3: Compose your request Take a glance at the character preview Enter your scene description in the prompt field Explore the prompt helper tabs Include keywords via the suggestion chips Pick how many you want generated Hit the main Generate Image button
4. Step 4: Check your creation Sit and wait for the process to finish Take a look at the result, if necessary, edit the text Hit generate again, or save

Recommendations for Optimal Quality

Keep it simple,

Resist the urge to pack as many details as possible into the first try.

Try beginning with a basic idea: sitting on a sofa, facing camera

Next, layer in extra specifics:

sitting on a leather sofa, sporting a crop top, gazing at camera, soft lighting

This typically yields better results than piling on a dozen attributes in one go, hoping the AI understands what you’re thinking.

Apply the prompt helper tabs

If you’re at a loss for ideas, give them a shot:

• Outfit
• Action
• Pose
• Background

These buttons are there for a reason. Use the system to your advantage.

Focus on a single character

To ensure multiple clips are consistent, reuse the same character card. That approach helps you maintain a consistent look throughout.

Pick the lowest possible batch count

Because 1 is chosen in the screenshot, I’d recommend starting with this minimum. Experiment with that first; adjust up if required.

Modify a single parameter per attempt

When the image isn’t coming out right, focus on changing just one variable to see what shifts:

• garb
• position
• environment
• expression

It becomes much easier to track which change actually made a difference.

Feature Reference Table

Conclusion

In conclusion, judging by the images, this generator features an easy-to-understand process:

1. first gain access,
2. then pick the subject matter,
3. after that describe what to create,
4. and finally generate and review.

This is a good approach, allowing a user to stay comfortable with simple steps while providing a lot of customizability through prompts, filters, and suggestion tabs.

For accuracy reasons, note that the currently visible text still says Generate Image, not Generate Video, so this could mean the platform is using a common generation interface. Since the end screen is not depicted, the last step is shown as a logical review process rather than an actual layout.