OneStar asks each review in a sample the same fixed questions with Jev, a model built for typed judgments: is there a complaint, which theme, is the reviewer leaving, are they asking for something. Code counts the answers into shares, trends and audience splits that read the same way for every app.
Every share opens the reviews it was counted from. When Jev is not confident, the review lands in an unsure row that stays on screen, so a report shows how much it does not know.
Connect App Store Connect or the Play Console and the reply inbox shows each review with the theme, severity and leaving signal Jev found, sorted so the people about to go come first. Your saved replies are offered where their stars and themes match, and you write or pick every reply that goes out.
On Pro and Max you can connect Claude, Codex or Cursor to your account. Your assistant reads the same reports and the reviews behind them, keeps notes on a report, and can start a report, a refresh, an Ask, a lens or a custom category your own model drafted, or track a report. Before anything starts, OneStar returns an estimate in reviews for you to approve, and Jev still judges each review.
Any report can go out as a read only link, and on Pro it can be published to the public gallery, where anyone can read it without an account.
OneStar does not write replies with AI or send automatic ones, has no Microsoft Teams integration, does not translate reviews and gives one seat per plan. If those are the job, Appbot is the better tool.
