Skip to main content
Every scheduled AI job in the business is listed in one place with its last run and the model it used, plus a button to run it again right now. Schedules and models change without touching code.

What ports as-is

  • A start-plus-run-id contract with a concurrency guard
  • Startup sweep that recovers a job left running by a crash
  • Per-job schedule override, re-read and woken on save
  • Self-registration of an external routine through its first run
  • URL-resumable run watching with a polling fallback

What we build for you

  • Which jobs and routines exist and what they write
  • Schedule labels and defaults
  • Which triggers count as event-driven