Reyn

Reyn is a Mac desktop application that monitors screen activity continuously and builds a local journal of your work. The core loop: Reyn watches what’s on your screen, indexes it on-device, and lets you ask natural-language questions about your own history — ‘what did we agree on in Tuesday’s call?’, ‘what was I looking at before lunch?’ — and returns answers grounded in that screen record. It also knows what’s on your screen at query time, not just what was there in the past, which means context for a question is live rather than stale. A morning email digest summarizes what you completed, what remains open, and what deserves attention — delivered without requiring you to open the app.

The differentiating claim is local-first data handling. The vendor states that screen data never leaves your Mac. AI inference runs through models you already have access to — the page describes multi-provider support — so the architectural split is: raw screen data stays local, AI calls go out to whichever model provider you configure. For users where screen content includes sensitive client data, code, or internal documents, this is the meaningful separation that cloud-based activity trackers don’t offer.

Where Reyn fits cleanly: solo Mac users who context-switch heavily and want a queryable record of their own week, or individuals who run repeatable processes and want those steps documented without writing them manually. Where it breaks: the tool is Mac-only with no self-hosted or Windows option described on the vendor page, no API is available, and the data model appears to be single-device. A team trying to share workflow documentation captured by Reyn, or a user who splits time between Mac and Windows, runs into a wall the product does not address. The Reyn Experts feature — a network of vetted AI engineers matched to frontier AI company roles — appears to be a separate service offering on the same domain, not a feature of the desktop tool itself.