HologemHolara Usage Guide

Creating Images

Holara is an app that converts a textual description of an image into amazing anime artwork.

View Demo Video


To create an image, go to the Create page and type a description of the image you want to create, then click the button.


You can also use the button to have a great image description automatically written for you.


Lastly, you can convert existing images including those made with Holara to new images based on a prompt. Check out the ImageImage section to learn more.



To learn how to write good image descriptions, called 'prompts', see Writing Prompts


To learn about image generation options, see Image Options


For more help and fun, check out our Discord and Twitter

Holara-sama!

How to Make Amazing Images

Use the button under the prompt box to help find amazing tags to use for your prompt!

Try including a tag from each category for the best results!

Writing Prompts

Prompting

Anything written in the prompt box will be converted to an image when you press generate. Writing prompts that result in the image you want to create is a skill that can be perfected over time.


Prompt Tags

For the best results, prompt using image tags whenever possible. For example, "blonde hair, red eyes, 1girl, solo" will perform better than "a girl with blonde hair and red eyes".

To find out more about tags, see Tags and Suggestions

Prompt Weighing

You can modify the weight of parts of your prompt so that they are more or less present in your output images:

  • To increase the weight of a word in the prompt, add a + after it: "girl+"
  • To decrease the weight of a word in the prompt, add a - after it: "snow-"
  • To modify the weight of more than one word, add parenthesis to the phrase: "(red hair)+"
  • To modify the weight of a word or phrase even more, you may add more + or - symbols: "(blue eyes)++"

Tags and Suggestions

Using tags in your prompt is a great way to create amazing images!

To help you find the right tags to use, tag completions are suggested under the prompt box as you type.

You can click on a tag completion to add it to your box, or simply finish typing the tag.

Hologems

Image Size Hologem Hologem Cost
512x512 Hologem4
512x768, 786x512 Hologem6
768x768 Hologem8
768x1152, 1152x768 Hologem12
1152x1152 Hologem20

All image creations cost Hologem hologems, which can be obtained from account creation, purchases, gift codes, and using Holara daily.

Image cost in hologems increases as the size of the image increases.

You must Hologem Subscribe To A Plan to purchase hologems.

After you are subscribed, you can purchase more hologems via the HologemAdd More button on the create page, or from your account page.

Image Options

There are many options to use when creating an image! Here's what they do.

Prompt: Text that will be used to create an image.

Negative Prompt: Can be used to attempt to exclude things from your images.


Model: Image model to generate an image with (currently only one model is supported).

Size: Size of the image to create.

Images: Number of images to make at once. Each image costs the same amount of HologemHologems.

ImageImage

Holara supports image-to-image functionality, allowing you to use any image you have as a starting point for your work.

Begin by clicking the ImageImage button, and then drag and drop an image into the space below it.

Image Strength

You may customize the image strength while using the image-to-image feature.


The lower the strength is, the more your creation will look like your input image.
The higher the strength is, the more your creation will resemble your prompt.

Image Variations

The image variations feature allows you to quickly use image-to-image functionality to generate variations of an existing image that you have already created.

Click on the 'variations' button on the bottom left corner of an image you have created to try it out!

If you'd like to customize your image variation settings before generation, you may turn off automatic variation generation under Advanced Options

Keybinds

Holara has keybinds to help you create images more quickly!

Keybinds while typing a prompt

While typing a prompt, you can hit Escape to stop typing, Tab to automatically type the first tag suggestion, and Enter to start generating your image.

  • Escape: Unfocus prompt / Stop typing
  • Tab: Use the first tag suggestion displayed
  • Enter: Generate Image

Other Keybinds

While not typing (which you can hit Escape for do), you can use the numbers 1, 2, and 4 to set the desired number of images to create, G to generate an image, V to create image variations, R to generate a random image, C to clear your prompt, P to start typing your prompt, and N to start typing your negative prompt.

  • 1, 2, 4: Set number of images to create
  • G: Generate Image
  • V: Image Variations
  • R: Generate Random Image
  • C: Clear prompt
  • P, N: Focus prompt or negative prompt

Advanced Options

There are many advanced options you can set as well. Here's an explanation of them!

Steps: Number of steps the image model uses when creating an image. Image detail increases with more steps, but will become overbearing at some point.

Scale: Scale determines to what extent the image will look like your prompt. Higher scales make the image more like what you type, but may also result in a lower quality image.

Seed: The seed is used in initializing randomness used to make your image. Use the same seed to get the same result with the same options and prompt!


Quality Tags: Holara may modify your prompt after you submit it to attempt to give you the best image possible. You can disable this if you'd like to have more control over your images.

Random Image/Prompt: This button toggles whether the 'Random' button will automatically generate a random image for you, or just suggest a random prompt for you to edit.

Variations : This button toggles whether the 'Variations' feature automatically creates variations for you with predefined settings, or whether it lets you select your setting prior to image creation.

Save Prompts: Your prompts can be saved to allow you to use the 'Favorite' button on images to favorite prompts, as well as to let us provide better suggestions to you for future prompts.If you disable this, no prompts you submit will be saved by us in any way, and you will not be able to favorite images or prompts.

API

Holara now offers an experimental API endpoint for image creation.
This endpoint is only available to subscribers. To use the API:

1. Generate an API key on your account page.
Do not share this API key - anyone with this key may generate images with your account, costing you hologems.
If you regenerate your API key, any previous API keys will no longer work.


2. Make a https post request to the API endpoint https://holara.ai/holara/api/external/1.0/generate_image with the following required parameters:

  • api_key (string) example: 'secret-99c34241-585d-4499-942a-be4e59b046d1'
  • model (string) example: 'Akasha: Anime'
  • prompt (string) example: '1girl, solo, black hair, red eyes, snowing, long hair, fox ears
  • num_images (integer, supported values: 1 and 2 for all image sizes, 4 for small and medium images) example: 1
  • width (integer, supported values: 512, 768, 1152) example: 512
  • height (integer, supported values: 512, 768, 1152) example: 512
Additionally, the following optional parameters are supported:
  • negative_prompt (string, optional) example: 'apples'
  • steps (integer, optional, supported values: 10-30) example: 28
  • cfg_scale (integer, optional, supported values: 3-20) example: 12
  • seed (integer, optional, supported values: 0-2147483648) example: 12



3. Ensure your request was successfully executed by ensuring the status code returned is 200. If a different status code is returned, there will be an error populated in the error field of the json response.


4. Use the returned base64 image data directly in a web application, or decode it and save it to a file


5. Keep the following in mind for any applications:
  • Image responses can be unpredictable, and Holara is not liable for what an API user does with created images. For more information, read our terms of service
  • API requests use the same hologems that your account has, and you will have to ensure you have a sufficient amount to create images. The amount of hologems an image costs can be found in the hologems table above
  • API requests are only available for users with an active subscription
  • API requests may take a long time, and in the event of a service outage may time out entirely, so be sure to set reasonable timeouts for your requests
  • API usage is still experimental and there may be bugs or other issues. Do not use this API for any critical services as it is not guaranteed to remain a feature yet. Feedback is welcome via Discord, Twitter, or Email
6. Here is an example Python script which will generate and save an image.
The script only requires that you set your API_KEY at the top. This is visible from your account page.

    import requests
    import json
    import base64

    API_KEY = 'secret-99c34241-585d-4499-942a-be4e59b046d1'

    url = 'https://holara.ai/holara/api/external/1.0/generate_image'
    data = {
    'api_key': API_KEY,
    'model': 'Akasha: Anime', #required
    'num_images': 1, #required
    'prompt': '1girl, solo, black hair, red eyes, snowing, long hair, fox ears', #required
    'negative_prompt': '', #optional
    'width': 512, #required
    'height': 512, #required
    'steps': 28, # optional
    'cfg_scale': 12, #optional
    }

    response = requests.request('POST', url, data=data)
    if response.status_code != 200:
    print(f'Error: {response.status_code} {response.content}')
    else:
    response = json.loads(response.content)
    # print basic response information
    print(f'Status: {response["status"]}')
    print(f'Execution Time: {round(response["execution_time"], 2)}s')
    print(f'Generation Cost: {response["generation_cost"]}')
    print(f'Hologems Remaining: {response["hologems_remaining"]}')

    # if you'd like to display the image on a website, you may directly set an image element's src to the base64 string
    # if you'd like to instead save the image, you must convert the base64 string to bytes and save it to a file:
    image_base64 = response['images'][0]
    image = base64.b64decode(image_base64)
    with open('image.png', 'wb') as f:
    f.write(image)

FAQ

How does Holara work?

Holara uses an AI model to generate images from text. This model is a computational process that uses diffusion. The goal of a diffusion model is to learn the underlying distributions of data so that it can then generate more samples similar to this data.

Although many of the processes involved sound complex, the basic idea is to find correlations between image descriptions and the images themselves, and to learn them so well that you can recreate new images just from their descriptions.

How is Holara similar to other AI image generation models?

Holara's user interface was made from scratch with simplicity, ease of use, and aesthetics in mind.

Holara currently offers two different models to create images with:

Akasha: Holara's first anime image model, Akasha is good at creating anime artwork, and works best when prompted with tags rather than natural language. It has been fine-tuned from the base model of Stable Diffusion 1.5 and does not use the weights of any other anime models.

Aika: Holara's second anime image model, Aika is similar to Akasha but offers more aesthetic, artsy, and creative outputs, even for simple prompts. Aika is based off of Akasha and is generally preferred by most users.

Can I use images created with Holara for commercial purposes?

Holara makes no claims over copyright or ownership on any of the images produced by the service, however also does not assume any liabilities. For detailed information, please see our Terms of Service.

Support And More

For more information or support, you can visit us on Discord or Twitter

For billing assistance or help with your account: [email protected]

For anything else, such as business inquiries: [email protected]