MetaRozetta's highest-accuracy (fully automated) mode is now available via API, allowing external systems to call the service directly to perform translations. Automatic integration of Japanese-English translation is expected to advance the automation of business workflows.
ロゼッタ, a consolidated subsidiary of メタリアル, announced that the highest-accuracy mode of its fully automated translation AI, MetaRozetta (formerly T-4OO), is now available via API. Text and file translations can now be called through the API, allowing translation processing and subsequent operations to be integrated directly into companies' business systems without human intervention. Supported language directions are Japanese to English and English to Japanese, and the service has been available since 5 October 2026.
The update is expected to support uses that incorporate translation into business processes, including automatically creating English versions of manuals and automatically translating specifications and internal rules from existing systems, as well as embedding the service in automation tools such as RPA to process large volumes of translations on a scheduled basis. Because translation results can be registered and distributed directly through internal systems, the scope of automation can be expanded while minimising the risk of mistranslation, according to the company.
As a long-term medium-term target, the company has set consolidated revenue of ¥15 billion for the fiscal year ending February 2030 and plans to address listing maintenance criteria with a view to meeting market capitalisation requirements. The current share price is ¥463 and the PER is approximately 31 times. The deterioration in first-quarter results was attributed to delays in large projects and the impact of upfront costs, while future revenue growth and margin improvement are viewed as key to a reassessment of the share price.
Facts':[
k01
Business partners
The fact that MetaRozetta can now be integrated directly with companies' business systems through API support is cited.