– Built our own AI engine that has learned from over 30 million templates, design sources, and user search history.
– Start of development of design-specific LLM, full-scale entry into the global design AI market
Design platform company MiriD Co., Ltd. (CEO Kang Chang-seok) announced on the 22nd that it has released 'MiricleNet', a design-specific AI engine that understands and analyzes users' design purposes and flow for the first time in Korea.
Miricnet is a design-only AI engine developed independently by learning from over 30 million templates, design sources, and user search history owned by Miric. Its features include recommendation and search functions that understand the user’s design context, such as the purpose, layout, and mood of the content, and find the optimal image.
The core functions of Miriknet are ▲ 'template search' based on multiple keywords ▲ 'style recommendation' based on design context. Template search recommends templates with the most similar layout and composition in real time when multiple keywords such as 'invitation, green, calligraphy' are entered, and the content component search function such as table of contents, chart, and SWOT analysis will also be expanded within the year. Style recommendation automatically selects and presents suitable illustrations, icons, and patterns from tens of millions of design assets without a separate search.
In particular, MiricleNet has also been applied to Miricle, an image creation and editing AI service launched earlier this year. Through this, users of Miricanvas and Bizhouse operated by Miridi can quickly and accurately search for references necessary for their work, thereby further increasing design work efficiency.
In addition to strengthening the independent design AI system centered around MiricNet, Miridi also plans to develop a design-specific LLM (large-scale language model) that can interpret design contexts more deeply in the future.
MiricleNet will be distributed simultaneously in the US, Japan, Europe, and other countries including Korea, and Miridi plans to accelerate its expansion into the global market based on design innovation that combines technological advancement and user experience.
Miridi Jang-shion, an AI researcher, said, “While existing generative AI was limited to content creation, the difference with MiricleNet is that it was designed to understand the design process and perform user-tailored judgments,” adding, “We will create a design AI ecosystem that anyone can use intuitively.”
You must be logged in to post a comment.