What problem does it solve?
It prevents afternoon from starting cold by detecting task drift and stale work, checking meeting records for gaps, and pre-loading relevant context so the user only gets nudged when something truly matters.
Core Features & Use Cases
- Pattern detection with delta state: Scans DuckDB task data and recent meeting records, using a persisted state file to surface only new or changed findings.
- Context warming for upcoming meetings: Pulls today’s remaining Google Calendar events and correlates external participants with prior meeting records and open threads.
- AI continuity checkpoint: Creates a lightweight midday memory breadcrumb (when Supermemory is available) to keep the afternoon aligned with earlier activity.
- Threshold-based delivery: Texts only for high-severity issues (e.g., 7+ days overdue), time-sensitive upcoming meetings, or genuinely new, important changes—otherwise stays silent.
Quick Start
Ask your scheduled agent to run midday-checkin at noon in the user’s timezone to deliver any urgent drift updates and pre-warm context for the afternoon.