This document introduces the features of Karlo APIs.
The Karlo API provides the ability to create new images based on sentences and images input by the user. Karlo, having learned from a scale of 300 million image-text pairs, understands the content described by the user and quickly generates completely new images at the pixel level. It also supports objects, backgrounds, lighting, composition, and various styles to enable creative activities according to the concept desired by the user.
Imagine how Karlo will transform your creativity and your brand's communications. Check out the features provided by Karlo.
As of February 22, 2024, the Karlo model used in the Karlo API has been upgraded to version 2.1 with enhanced image generation quality.
Prompt is an input text to generate an image through Karlo API. Prompts can be keywords, sentences (in English), or images that describe the image you want to generate.
Karlo API supports only English.
Below is a sample of what images Karlo will generate with a given prompt, depending on the model version. Karlo 2.1 can generate higher resolution and higher quality images compared to the version 2.0.
A photo of a cute tiny monster on the beach, daylight
You can also generate images with various compositions or styles. To derive a better result, modify and adjust the prompt. Refer to How to use to check out tips and examples using the prompt effectively.
Negative prompt allows you to set information to be excluded from generated images through Karlo API. You can enter keywords and sentences to exclude elements that interfere with image generation.
Below is an example of using the prompt A dragon wrapped in blue fire
to generate an image, and using the negative prompt Side view
to make the subject face forward in the image. You can see how to effectively configure prompts and negative prompts with examples in How to use.
Karlo creates new images based on the given text prompt. You can also specify a season, an author's style, or a graphic style.
You can generate images with a size of up to 1024 pixels in width and height. A sample is below.
Create variations of the given image. Karlo understands the context of the given image and regenerates a new image with another composition or characteristic.
A sample is below.
Creates a connected image based on the given image by understanding the contents of a selection or boundaries. Reconstruct a selection, or create a continuous image outside the image boundaries. You can modify or expand an image while maintaining the structure and style of the original.
This feature is only available with Karlo model version 2.0.
A sample is below.
Companies or customers ("users") who utilize Karlo API provided by Kakao Brain ("Brain") through Kakao Developers, must agree to the Terms of Use, in addition to consenting to the Terms of Service and Operating Policy of Kakao Developers.
Karlo API provided on Kakao Developers has quotas (usage limits) on a per-minute, daily, and monthly basis to ensure smooth service delivery. The currently applicable quota information can be found in the Quota. For an upward adjustment of the applied quota limit, inquiries, and partnerships, contact us through the partnership portal.
Feature | Description | Documentation |
---|---|---|
Generate image | Creates new images based on the given text prompt. | REST API |
Make variation | Creates variations of the given image. | REST API |
Modify image | Creates a connected image based on the given image. | REST API |