[AI Revolution]⑧ Turning Escape Zebra 'Sero' into a Superstar... Ryan Rocket

Ryan Rocket's AI Image Generation Web Platform 'Sporky'
Zebra 'Sero' Various Parody Images Generated 1,250 in One Day

An image parodying the zebra 'Sero,' who recently made headlines for escaping from Children's Grand Park. [Image source=Sporky]

Recently, a zebra escaped from Children's Grand Park and roamed around downtown Seoul for about four hours. Many citizens showed great interest by sharing the real-time movement route of this male zebra named 'Sero' on social networking services (SNS). The majority of emotions expressed in videos featuring Sero or in comments on related articles were more of sympathy than fear. Numerous posts offering comfort resonated widely, suggesting how frustrated a two-year-old, who should be freely running on green meadows, must have been to escape from there.

Sero Reborn in RyanRocket's 'Sporky'

Perhaps because of the collective desire to grant Sero freedom, he has now become a superstar in the virtual world. This is through 'Sporky,' an image generation web platform by AI startup RyanRocket. Sporky is a service that instantly provides images based on text descriptions input by users. For example, if you enter 'a zebra in an astronaut suit with the moon as the background,' images matching this description will appear. Users can also choose various formats such as photos, illustrations, or animations. Starting with some netizens creating and sharing parody images on Sporky to fulfill Sero's dream, various images have now spread like memes (internet viral content).

When you access Sporky and enter 'Sero's dream,' thousands of parody images are found. These include Sero dancing on stage, flying with wings, and striding upright on two legs through the streets. Just one day after the Sero escape incident on the 23rd, over 1,250 Sero-related images were created. Jung Seung-hwan, CEO of RyanRocket, said, "This incident confirmed that generative AI is gradually entering everyday life, as netizens voluntarily use image generation AI to create various parodies."

More than 1,250 images related to vertical orientation were generated in one day by domestic and international netizens active on the image generation artificial intelligence (AI) workflow web platform 'Sporky'. [Image source=Sporky].

Founded to Help the Visually Impaired... Focused on Advancing Generative AI Technology

Founded in 2019, RyanRocket aimed from the start to advance generative AI technology. CEO Jung, who majored in Information Systems at Hanyang University, came across a YouTube video during his university years where a visually impaired child expressed a desire to listen to storybooks freely. Along with two colleagues, Jung studied voice synthesis technology and developed voice generation technology in 2018 to create audiobooks for the visually impaired. The following year, they started the company with four employees. Jung said, "Audiobooks require significant costs and time for voice actor casting and recording, and their prices are high, making it difficult for visually impaired children to access them. Our goal was to produce hundreds of thousands or even millions of storybooks for them."

Jung Seung-hwan, CEO of Ryan Rocket.

In its early days, RyanRocket secured seed investment from Mashup Angels, a startup-focused investment firm. It then received Series A investments from ID Ventures, Daesung Startup Investment, Harang Technology Investment, and IBK Industrial Bank of Korea. The total accumulated investment to date is 7.7 billion KRW. In 2019, the company was also selected for TIPS, a technology startup support program by the Ministry of SMEs and Startups. RyanRocket developed services capable of creating various types of content, including images and videos, starting from text-based voice generation technology.

Besides Sporky, RyanRocket operates OnAir Studio and VeryMe, platforms for creating virtual human content. OnAir Studio is a platform where AI completes dubbing, subtitles, and video production all at once with simple operations like text input, without complicated video production programs. VeryMe is a 'face forming' service that creates a personalized virtual face based on the actual user's face. It can recognize not only still faces in photos but also moving faces in videos and convert them into virtual humans. RyanRocket plans to launch VeryMe in the global market by the end of this month.

"Surge in Inquiries About Generative AI... Providing Industry-Specific Content"

CEO Jung said that with the recent rise in interest in generative AI due to ChatGPT, inquiries have flooded in from various industries such as webtoons, education, and interior design. As an early domestic generative AI startup, many seek to learn from or propose collaborations based on the accumulated know-how.

Jung has set a goal to fine-tune ultra-large generative AI models to provide customized content tailored to each industry. Fine-tuning refers to adjusting services in detail according to the purposes of individual companies or users. Jung said, "In webtoons, there are cases where specific art styles, compositions, or styles are desired. We will advance generative AI technology to allow such detailed adjustments." He added, "As various possibilities open in the generative AI era, our business roadmap for this year is to evolve existing services into ones that provide greater value."

An image generated by entering a command on 'Sporky,' the image creation web platform of the AI startup LionRocket.