Reference

Pages and syncing

JSON lists and reports come in pages. While meta.next_cursor isn't null, call the same URL with cursor set to it; meta.total gives the number of rows across all pages. A CSV is a whole file: without limit or cursor it holds every row (up to 100,000), and a bigger one is refused rather than cut. Add limit to page a CSV too; the next cursor is then in the X-Next-Cursor header. The pages of one report or feed read come from the same moment: the first page keeps the rows for 15 minutes (meta.as_of tells when they were read), so the next pages answer quickly and never shift when a planning changes in between. A cursor used later still works and reads fresh rows.

"meta": {
  "count": 500,
  "total": 1240,
  "limit": 500,
  "next_cursor": "odw"
}

Keeping a copy in sync

Store the time of each sync. Next time, call /plannings with changed_since set to it, then re-read the weeks it returns. A week that keeps changing after publication shows up again.

GET /api/v1/plannings?changed_since=2026-10-12T08:00:00Z

# for each planning returned:
GET /api/v1/reports/agent-days?activity={activity.code}&week={week}
Pages and syncing — Planifizer API docs