Work out what is due and when to run next (schedules)
Scheduler: from recurring schedules (every day or week at a local time in a time zone, daylight saving handled) and one-off timed actions, work out what is due to run now and the next time to run
Give now plus schedules (schedule_id, timezone as an IANA name, start_local_date, cadence {unit: day | week, interval?, weekdays?: 0=Sunday..6}, local_window {start: HH:MM, duration_minutes}, dst {gap: skip | next-valid, fold: earlier | later | both}, action with its owner_dispatch) and their occurrence_states, and/or one-off time_windows, actions and receipts. Returns recurring_due, scheduled_due, the occurrences to record before running them, and next_wake_at: the next time anything is due, counting what is due now as done.
What it asks for
- now (Needed, value): The time to work from: ISO text or epoch milliseconds
- schedules (Optional, list): Recurring schedules
- occurrence_states (Optional, list): What was recorded for earlier occurrences (only a proven owner_outcome moves a schedule on)
- time_windows (Optional, list): One-off times: {record_id, start}
- actions (Optional, list): One-off actions: {record_id, action_id, dispatch: {name, input}}
- receipts (Optional, list): One-off actions already done: {record_id, operation}
- max_calendar_scan_days (Optional, number):
Use it from your AI
Connect the AI you already use to Trillion once. Then ask in your own words; your AI finds this tool by what it does and runs it.
Say to your AI: “Use Trillion to scheduler: from recurring schedules (every day or week at a local time in a time zone, daylight saving handled) and one-off timed actions, work out what is due to run now and the next time to run”
What your AI calls: run_tool {"tool_id":"tool:when-to-run-next"}