
About the Project
In the future Techno_Underworld, a girl confronts her inner fears through a ritual of sound and light, facing monsters born from memory—Ha (chaos), Lei (storm), Lu (disorientation), and Ya (panic).
In the Techno_Underworld of the future, a girl lives in a world where tradition and technology intertwine. Each year, villagers hang red lanterns to ward off evil, yet the girl begins to question whether rituals can truly resist future threats. One day, she discovers a wolf carrying a red lantern connected by a red audio cable—a signal of the flow of life and time. When she brings the lantern home and reignites the failing energy, time freezes; the "Future Beads" device has run out of power. From a virtual rift, four monsters emerge—Ha (chaos), Lei (storm), Lu (disorientation), and Ya (panic)—manifestations of her inner fears. Battling alongside the villagers through deafening sound, she ultimately chooses to accept her fears. As understanding arises, the monsters vanish. In reality, the girl had already died of oxygen deprivation; the journey was an AI reconstruction of her final memories. Her story becomes an eternal resonance within the Techno_Underworld, carrying courage, wisdom, and release.
AI Tools & Workflow
This project utilized several AI tools across different creative stages. MidJourney was employed primarily for visual ideation and reference imagery, helping to conceptualize the world-building elements such as atmosphere, textures, and symbolic objects like the red lantern. Runway and HailuoAI served as primary tools for animation generation and scene transitions, allowing dynamic visual storytelling. Sora contributed additional support in generating motion sequences and enhancing environmental details. AI tools not only enabled me to bring abstract ideas into tangible form but also accelerated my creative workflow. They provided a foundation for experimentation, strengthened my confidence in visualizing complex narratives, and allowed me to refine techniques more efficiently than traditional methods.