I developed Pace to help students return to schoolwork after a disruption. I designed it so that you can plan your work without explaining what happened.
I designed Pace to store your tasks, steps, check-ins, plans, focus sessions, reflections, and settings in your browser's local storage (IndexedDB) on your device. There is no account and no server-side database. You can export everything as a JSON file, import it on another device, or delete it all from Settings at any time.
I use a language model in Pace to write your daily plan, break tasks into steps, read pasted text or photos, and draft teacher messages. I route these requests through Pace's own serverless function to the model provider. Each request contains only what that feature needs:
I exclude private task notes, your profile name, and raw reflection records from planning requests. Text, photos, and names you explicitly submit to Capture, task breakdown, or teacher-message drafting are sent as described above.
I keep Pace's serverless function stateless: it forwards the request and streams the answer back without keeping a copy or logging the content. Every request asks the provider not to keep a retrievable copy.
If the planner cannot be reached, Pace plans entirely on your device with its built-in scheduler, and nothing leaves your browser.
If you connect a Canvas or Google Calendar feed link, Pace fetches it through its serverless function (browsers block the direct read) and parses it on your device. The function does not store the link or the calendar. The link is kept in your browser's local storage; treat it like a password, since anyone with it can read that calendar.
I do not use advertising, third-party scripts, or cookies in Pace. Fonts are served from Pace itself. The anonymous device token Pace generates is random, is not tied to your identity, and is used only to rate-limit the public planner.
You can contact me at sohanm4364@gmail.com. I have made the code open source so that the implementation can be checked against this policy.