Revieko is a pull-request review aid focused on structural changes in a codebase. Its homepage describes learning a repository-specific baseline from history and highlighting places where a new diff departs from those patterns. The intended output is an attention map for human reviewers: a short risk summary, line-level hotspots and per-file actions. It explicitly distinguishes this role from replacing code review, finding every bug or enforcing cosmetic style rules.
The workflow begins by installing the GitHub App for selected repositories. Revieko describes building the baseline automatically and returning an evolving pull-request comment and status check. Its default is to warn rather than block, with policy controls described as configurable. For deeper inspection it lists reports in HTML, Markdown and JSON, plus a glossary explaining the drift categories. The public examples include new abstractions, hidden coupling, introduced state, dense control flow and structural pattern breaks.
The site characterizes its approach as structural machine learning rather than language-model interpretation of code as prose. A free core offering is listed, while team and enterprise options are described as pilots. Developers should review the actual app permissions, repository access and current plan before connecting production code. A hotspot score is a prompt to investigate, not proof that a change is wrong. Revieko fits teams that want help choosing where to begin a large review, while tests, security checks and engineering judgment remain separate responsibilities.
