rex0220 kSQL Dashboard Pro - AI(Claude Code)にダッシュボードを作らせるkSQL Dashboard Pro lets developers describe kintone dashboard requirements in…

kSQL Dashboard Pro is a newly published environment that lets developers create dashboards for kintone, the Japanese business application platform, by describing their requirements in plain Japanese and letting an AI coding assistant such as Claude Code generate the implementation. For teams that build operational apps on kintone, the approach aims to cut down the manual effort normally involved in assembling aggregations, period filters, and screen layouts.

The tool comes from rex0220, a developer who has produced a range of kintone-focused utilities, and was documented on Qiita. It builds on kSQL, a query layer that brings SQL-style access to kintone data. The "Pro" dashboard release appears to extend that foundation so that an AI agent can turn written specifications into functioning dashboards rather than requiring the developer to wire up each component by hand.

The workflow centers on a single natural-language instruction. In the example given, a user asks the AI to build a period-switchable dashboard on the "【全社】ダッシュボード" (company-wide dashboard) list within app 4149, described as a case management app. The instruction specifies that the period control should offer options such as this month, last month, this year, last year, and a custom date range. From that description alone, the system is intended to produce the dashboard, including the dynamic period filter and the layout implied by the requirements.

That emphasis on period controls is notable because time-based filtering is one of the more repetitive and error-prone parts of dashboard building. Handling relative ranges like "this month" versus "last year," and letting users toggle between them at runtime, usually requires careful date arithmetic and state management. Delegating this to an AI that already understands the kSQL query model could remove a common source of tedium, though results in practice will likely depend on how precisely the requirements are phrased.

For readers unfamiliar with the underlying pieces, kintone is a low-code platform from Cybozu that lets organizations build database-driven business apps with configurable lists, forms, and views. It supports customization through JavaScript and third-party plugins, and dashboards are one area where teams frequently reach for such extensions. kSQL fits into this ecosystem by exposing kintone records through a familiar query syntax, which makes aggregation and reporting more approachable than the platform's native tools alone.

Claude Code, the AI assistant named in the announcement, is Anthropic's agentic coding tool. It can read a project's files, run commands, and generate or edit code, and it is often used inside editor environments; the tags accompanying the post reference VSCode, suggesting the intended setup sits within that kind of development context. The combination reflects a wider industry pattern often labeled AI code generation, in which developers describe intent in natural language and let a model produce the corresponding code.

kSQL Dashboard Pro is one example of that pattern applied to a specific, practical domain rather than general-purpose programming. Similar efforts have appeared across the low-code and business-tooling space, where vendors and independent developers are experimenting with natural-language interfaces layered on top of existing platforms. The appeal is that domain-specific context, in this case knowledge of kintone apps, fields, and the kSQL model, can make an AI's output more reliable than a generic prompt would allow.

Several practical considerations remain worth noting. The examples are presented in Japanese, and it is unclear how the environment behaves with instructions in other languages, though the technique is not inherently language-bound. As with any AI-generated code, generated dashboards would benefit from review, particularly around date logic and access to company-wide data. The published material reads as a hands-on demonstration of the workflow rather than a claim of fully autonomous dashboard creation, so teams evaluating it should expect to iterate on prompts and verify output against their own data.

For kintone developers who spend significant time on reporting views, the tool represents a concrete attempt to shift dashboard work from manual configuration toward describing outcomes in ever