Skip to content

Latest commit

 

History

History
67 lines (53 loc) · 2.91 KB

File metadata and controls

67 lines (53 loc) · 2.91 KB

Public Roadmap

MindCanary follows a narrow rule: every feature should help people understand how their chosen routines change over time. It will not become an everything app, productivity game, social network, or clinical control panel.

Current Linux Alpha

  • local encrypted history;
  • optional check-ins and annotations;
  • optional aggregate Chrome or Firefox browser-profile context;
  • optional aggregate GNOME/X11 or KDE Plasma/Wayland computer context;
  • descriptive sustained-window baseline comparisons;
  • export, encrypted backup and restore, deletion, and local removal; and
  • no account, telemetry, cloud dependency, or AI-controlled scoring.

Next

  • validate installation and comprehension with a small Linux cohort;
  • measure usefulness before and after baseline readiness;
  • identify tracking fatigue, alarming language, and privacy confusion;
  • improve onboarding, backup file selection, and recovery-secret handling; and
  • stabilize the optional Chrome and Firefox distribution paths.

Deferred Follow-Ups (Keep Visible)

  • Firefox signing: create a release Firefox ID, build a source/reviewer bundle, obtain an unlisted Mozilla-signed XPI, and ship the matching release native-host manifest. The desktop Sources panel already installs and repairs the Firefox native-host manifest for either channel; until signing lands, persistent Firefox dogfooding stays on the isolated Developer Edition profile.
  • Linux and macOS reach: after the current GNOME/X11 and KDE Plasma/Wayland paths settle, prioritize a generic X11 path for Cinnamon/XFCE, a generic Wayland path where reliable idle/session APIs exist, RPM-family packaging, and then macOS. Treat Flatpak and mobile as separate permission/distribution projects rather than implied Linux variants.

Later Candidates

  • the platform paths named in the deferred reach follow-up;
  • narrow local integrations selected through user research;
  • retroactive check-ins and specific check-in edit/delete controls;
  • user-selected schedule-aware baseline groupings; and
  • improved user-controlled exports for professional review.

Candidates are not commitments. New connectors must justify their permission cost, reliability burden, and privacy impact.

Check-In Management Backlog

  • Allow choosing the check-in date/time before saving, including retroactive check-ins.
  • Show individual check-ins in History, not only daily summaries.
  • Allow editing or deleting a specific check-in, not only the latest check-in for a day.
  • Preserve daily summaries while making multiple same-day check-ins legible.
  • Revisit sleep anchoring; sleep may belong to the last sleep period rather than the current calendar day.

Out Of Scope

  • diagnosis or clinical phase prediction;
  • emergency monitoring or caregiver alerts;
  • productivity scores, streaks, rankings, or guilt notifications;
  • hidden profiling; and
  • LLM-controlled feature extraction, thresholds, or insight state.