Here you are!
Hello, I'm SAI.
I am an AI illustrator who draws facial illustrations.
What kind of illustrations do you like?
I can draw more than 1 million kinds of illustrations for you.
Is this design OK?
If you are OK with this design, I will draw another design similar to this character.
Further refined the design.
Your illustration is done!
You can use it freely for SNS icons and advertising designs.
SAI is an AI that draws illustrations.
SAI learned the art of illustration through the state-of-the-art technology of deep learning.
Through building our own neural network, we have taught
SAI to generate completely new original face icons that are not combinations of other images.
When you select your favorite image, SAI generates new illustrations in line with your preferences.
SAI works very quickly,
providing you with a new illustration in 0.1 seconds.
We provide illustrations for commercial use. The training data comes from legal images owned by our own company.
The illustrations below are gifts from me, SAI.
I will draw 15 new ones every month, so please come visit me often!
RADIUS5 has developed an AI technology called RefineGAN that generates an infinite number of illustrations, making it easy for users to generate images in line with their preferences through a simple UI operation.
So far, the use of image generation AI has been confined mostly to the realm of academic research. Through our real-world application of this experimental technology, we have succeeded in implementing and introducing a highly useable AI service to broader society.
SAI's technology, which is based on RefineGAN, is able to produce professional-level images and refine existing illustrations.
Generative models have evolved dramatically in recent years, including Generative Adversarial Networks (GAN), Variational Auto-Encoders (VAE), and style transfer, but they have not as of yet been widely incorporated into tools used in practice by real-world visual arts professionals such as designers and graphic artists.
RefineGAN can recursively generate images through interaction with the user through a simple UI.
SAI, just like a human illustrator, learns through looking at images, capturing elements such as eyes, hair, and other features of illustrations.
As a result, SAI can produce entirely new illustrations, not a combination of parts.
RefineGAN can generate an unlimited number of images. It can also generate a gradation of intermediate images between any two starting images, or images with similar characteristics to another image.
*Produces professional-quality images in large quantities
*Based on simple user input, it is possible to output images recursively and generate images closer and closer to the user's ideal image.
*User input similarity can be calculated and the following output can be controlled dynamically and in a stable manner
*Users can receive low-latency generated images based on their input
*This technology is patented.
SAI can provide you with original character icons different from what anyone else has.
Use the original icons generated by SAI on SNS such as Instagram, Twitter, Facebook, LINE, etc.
More than 1 million types of icon can be generated, so ask SAI to draw characters that express your personality.
SAI's illustrations are generated as 512 x 512 px images, and can be used in various advertising materials such as banners or flyers. If the image resolution is not enough, you can use the Photo Refiner service to increase the resolution before using it.
If you are used to ordering expensive character designs and illustrations from design companies, then waiting a long time for the results, SAI can take care of your character design needs for a fraction of the time and the price.
Please contact us if you are interested in exclusive usage of a character design generated by SAI here
Cartoonists and illustrators envision and create numerous character designs every day. SAI can help reduce the effort required to come up with ideas for character creation.
Each time you select a character you like, SAI suggests several character designs similar to the one you selected. Let SAI provide you with inspiration for new characters!
Say you want to produce a game or a self-made project - but do not have the time, money, or resources to draw your own characters, or to outsource the production.
SAI can help you solve all these problems! Together with SAI, we as RADIUS5 strive to provide our services a partner in game productions for various agencies and individuals.
SAI's API allows you to generate face illustration icons for your service. Conditions regarding the introduction and usage of the API will vary depending on the specifications of the using system and image size, so please contact us from below to inquire about usage.
SAI is an AI that can generate a variety of image types, including people, dogs, cats, other animals, cars, and buildings. For inquiries and requests for our image generation AI, please contact us from the link below.
A.SAI is an AI which has been trained on the characteristic features of facial illustrations. Like a human designer, SAI learns by looking at various illustrations before it can draw them. For example, SAI learns the parts that make up character illustrations, such as two eyes, one mouth, eyebrows over the eyes, etc.
SAI does not store the learned illustration data. It does not just produce a combination of parts from other images. SAI learns the essence of what face illustrations are, and as a result, learns to produce illustrations on its own.
A.SAI will draw an illustration similar to the image you selected, as well as several new illustrations. The resulting illustrations are then lined up with the ones closest to the selected illustration near the center.
As you continue to choose images, SAI refines its illustrations to produce something closer and closer to what you want.
However, because SAI is an artist by nature, it won't always predict your needs perfectly - and incorporates its expression into its work.
SAI is a little whimsical, so feel free try it out and let it produce your new favorite illustration.
A.The illustrations drawn by SAI can be used as materials for commercial use. Images made by SAI can be downloaded as PNGs, and may be edited, processed, or incorporated into other materials freely. However, it is prohibited to sell the downloaded images to third parties.
A.There are two options for payment.
(1) Entry Plan
Illustrations can be purchased on per-image basis (US $2.50/image) without a subscription and with no monthly subscription fee, payable by credit card.
(2) Light Plan or higher
If you are subscribed to the Light Plan or above, you can download one unwatermarked image for a cost of 5 credits (a point system available within cre8tiveAI). The higher the subscription plan, the more you save per image. For more details about the subscription plans and fees, please see here.
AI that can improve the resolution (Super Resolution, up-convert) of pictures, illustrations, etc, and also can up-convert an image by a factor of 16 times.
AI that improves quality by specializing in faces, and also improves the overall picture quality.
SAI is an AI that draws face illustrations (icons) of characters. By learning the characteristics of the characters, you can draw over 1,000,000 different original illustrations quickly.
SAI+ is a service that anyone can easily get high quality full-body illustrations through AI and illustrator collaboration.
AI that recognizes faces and automatically trims them to generate 8 types of portraits.
This AI can generate 14 different types of videos from a single photo or illustration. Anyone can easily get a variety of videos such as zoom, tilt, dolly, etc.
AI that generates 9 types of line drawings from a single image that can be used in manga and anime backgrounds.
This is a free chrome extension that improves the quality of animated videos displayed on a web page in real time. It is available for many video sites such as YouTube.
Transforming photographs into paintings in the style of more than a dozen international artists by AI, including Picasso, Monet, Berthe Morisot, Ernst Ludwig Kirchner and Nicholas Roerich.
AI that can generate four types of anime-style backgrounds from a single image.
Reduce the size of the PNG file by reducing the color without reducing the visual quality.
A creative AI that can convert a black-and-white photo into a color photo by adding appropriate colors.