Hivemind Selected for “Busan-Style Unicorn” Candidate Program (Bunicorn Build)… Advancing Global Distribution Infrastructure for K-Content Through AI Localization

Be SUCCESS, April 7, 2026

AI localization platform company Hivemind Co., Ltd. (CEO Warren Kim) has been finally selected for the “2026 Bunicorn Build Development Program (Busan Early Startup Package)” hosted by the Busan Metropolitan City and the Busan Institute for Technology Entrepreneurship & Investment.

“Bunicorn” is a portmanteau of Busan and Unicorn, representing a Busan-specific brand for high-growth technology startups with proven potential. It was introduced by the city of Busan to strategically support such companies in scaling into unicorn enterprises. This year, Busan has placed “startup growth” as the top priority of its entrepreneurship support policy and established a step-by-step support system—Preliminary (SEED), Early Stage (BUILD), Growth (BOOST), and Busan-style Unicorn (BUNICORN). Through this framework, the city has built a growth-centered startup support model that spans the entire lifecycle, from discovery to scaling.

The Bunicorn Build Development Program (Busan Early Startup Package), to which Hivemind has been selected, targets technology startups with less than 1 to 3 years of operational history. Selected companies receive not only commercialization funding but also practical programs to strengthen investment capabilities, including consulting, mentoring, investor meetups, and connections with specialized institutions, along with follow-up networking support.

Hivemind’s core product, Aurorah, is an end-to-end localization SaaS platform supporting over 30 languages across text, documents, images, and video. It provides an integrated workflow powered by a single AI engine—from automatic subtitle generation to multilingual translation, editing, and distribution. Its primary clients include broadcasters, OTT platforms, and content creators. A key feature is its “long-context preservation” technology, which processes full-length web novels (typically around 1 million characters) within a single context, ensuring consistency in narrative flow, character voice, and settings. The platform operates on a collaborative model where AI performs initial translation, followed by professional translators who review and refine the output using a dedicated editor. According to Hivemind, the system achieves a Korean speech recognition Word Error Rate (WER) of 2.73% under KOLAS certification standards and a translation quality score of BERTScore F1 0.89—meeting the quality requirements of broadcast and OTT platforms.

Hivemind has signed web novel supply agreements with Korean publishers Insight Breeze and Storynlab, and is currently carrying out Japanese translation and localization of Korean web novels. Through partnerships with content companies including Eo Content Group, the company is expanding its applications into areas such as video IP and e-commerce image translation. Additionally, Hivemind received the highest rating (“Excellent”) in the “2025 Super Gap Startup 1000+ Project (MicroDIPS AX Consortium)” led by the Artificial Intelligence Industry Convergence Agency. The company has also demonstrated practical achievements such as simultaneous multilingual publishing for virtual creators and real-time subtitles and translation at live performance venues.
CEO Warren Kim stated, “Being recognized as ‘Excellent’ in the MicroDIPS evaluation and selected for Bunicorn Build is a meaningful achievement that validates Hivemind’s technological capabilities and commercialization potential at both national and regional levels. Starting with Japan, we will accelerate our global expansion so that Aurorah can become the standard infrastructure for K-content distribution—from individual creators to global OTT platforms.”

Through its selection for the Bunicorn Build program, Hivemind plans to strengthen its investment capabilities and leverage network support while preparing for the next stage of growth alongside its global market expansion.

Read Original »