Dimension scores are derived from public data and fields; weighted into the composite. Reference only.
datagusto is a technology startup based in Tokyo, Japan, founded in 2020 and positioned as a “manager for AI Agents.” Its core businesses include the AI Safety Platform “datagusto” and the education-focused AI Agent “fukutan.” The former targets reliability and safety issues in enterprise AI Agent operations, while the latter supports teaching and learning in schools, functioning somewhat like an “assistant homeroom teacher.”
Based on the official website, datagusto’s focus is not general-purpose chat or content generation, but the safe operation of AI Agents. The platform can monitor and evaluate AI Agent behavior in real time, and dynamically generate Guardrails to detect and prevent “unexpected” incidents that static rules may fail to cover. This positioning makes it suitable for companies that have already introduced autonomous agents into business workflows, especially scenarios that require preventing unauthorized actions, abnormal decisions, or uncontrollable automation.
fukutan is its education-oriented product. It can assist teachers with lesson design, assignment and task creation, and understanding students’ learning progress, with the goal of enabling both teachers and students to use AI with confidence. The website also mentions that its beta version is being tested in a proof-of-concept experiment with ICU, but it does not provide more detailed performance data.
The official website does not disclose its pricing model, plans, free quota, or trial entry point. It only provides contact options for inquiries and demos, so it appears to follow a customized enterprise or institutional sales model. API, SDK, and third-party system integration methods are also not specified. In terms of data privacy, the website emphasizes safety and trust, but we did not find concrete details on data storage, log retention, model training usage, access permissions, compliance certifications, or similar matters. These should be key questions before procurement.
The main advantage is its clear positioning: it directly addresses the risks of “loss of control” and unexpected behavior in real-world AI Agent deployment. The team’s background also appears relevant: the founder has experience in AI governance consulting, while the CTO has a background in data infrastructure and web data mining. The downside is that the publicly available information still feels early-stage, with limited details on technical architecture, evaluation metrics, false positive/false negative performance, performance overhead, or customer case studies. It is therefore difficult to judge the product’s maturity from the website alone.
It is better suited to enterprises that are deploying AI Agents and need safety evaluation and operational guardrails, as well as educational institutions willing to pilot AI tools in schools. The official website does not provide information on access from China, so this remains unknown; payment methods are also not disclosed. If alternatives are needed, it may be worth comparing Lakera Guard, NVIDIA NeMo Guardrails, Guardrails AI, LangSmith, Arize Phoenix, and other AI safety and observability tools, or building an in-house enterprise governance framework.
⚠ This review is compiled from public sources and does not constitute a purchase recommendation. Verify all facts on the vendor's official site. Verify on datagusto.ai official site.
datagusto.ai is an Japan AI Apps provider. TG4G tracks its product information, an overall rating of 7.0/10, and a China-accessibility score of Workable. Click "Visit Official Site" to reach datagusto.ai directly.