Enloop has no server and no accounts. It collects nothing, transmits nothing, and stores everything on your machine, in places you choose.
On sites you have granted access to, and only while you run a case there, the extension acts on the page: it highlights elements, inserts values you chose, and executes a case's automated steps. If you switch capture on for a run, the page's console output and request log are recorded into the local run files — those may contain data shown by the page, and they stay on your machine under your control.
Nothing, to us — Enloop's developers operate no backend and receive no data, no telemetry, no analytics, no crash reports. Sharing is always your own action: a share link carries the case inside the URL (the online viewer renders it in your browser; the case text in the link's fragment is not sent to the hosting server), a downloaded HTML page is a local file, and anything you commit to a repository goes where you point it.
The Claude Code and Codex skills run inside your own coding agent, under your account with that provider, subject to that provider's terms. Enloop adds no collection there either: the skills read your repository and write case files into your local data folder.
Changes to this policy are commits to the public repository — diffable, dated, and attributable, like everything else there.
Questions or concerns: open an issue at github.com/enloop-md/enloop/issues or write to ryabenko.sergey@gmail.com.