Tokyo Tech Firm Releases New Software to Secure Automated Artificial Intelligence Web Operations
Tokyo, Wednesday, 30 September 2026.
Launched on September 30, 2026, Life Scape’s Kagemusha isolates web access for autonomous artificial intelligence agents, boosting execution efficiency by nearly 40 percent while maintaining enterprise cybersecurity.
Enterprise Security and Application Architecture
Life Scape, K.K., a Tokyo-based software firm, officially launched Kagemusha on September 30, 2026, a macOS application designed to isolate artificial intelligence agent web access [1][3]. The software addresses enterprise security concerns by packaging individual websites into distinct application windows, restricting AI agents to designated signed-in web services [1]. Kagemusha’s architecture separates processes, logins, and control endpoints for each application, ensuring AI agents cannot access services outside their assigned scope [1]. The system utilizes browser symlinking to reference the user’s existing browser binary without modifying the host browser, allowing automatic updates while maintaining isolated profiles for cookies and logins [4]. The software includes presets for services such as SeaArt, Google AI Studio, and Duolingo, allowing users to convert them into standalone macOS apps [2][5][6].
Performance Metrics and Market Availability
Internal measurements by Life Scape indicate significant efficiency gains, with parallel task execution across three apps completing in 6.4 seconds compared to 10.2 seconds for sequential execution [1][3]. This performance improvement represents a time reduction of 37.255 percent in task completion time [1]. The software is available for free for up to three apps, while the Kagemusha Pro version costs $24.99 as a one-time purchase for unlimited app support [1]. Life Scape plans to release a Windows-compatible version by December 15, 2026, to expand market reach beyond the macOS ecosystem [3]. CEO Hisashi Hirajo stated that the company will continue strengthening the mechanisms for human and AI agent collaboration [3].