get jobs
Background work: generation, render, language change, imports, crop subject detection, file reads (transcript, search index, shot log, inspect_media prompt), and desktop app reads and uploads. With no ids: every job running, with its prompt, model, target item and who started it. With ids: those jobs, once all finish or wait_seconds passes. A waiting job has its queue place and eta_seconds left when past runs give one; cost_credits is its cost so far, a finished job’s cloud_minutes what it used. A finished generation, render or PDF has its item id, url and a link; a render missing a picture lists it in not_drawn; a finished language change carries timeline_id; a finished prompt read carries analysis and moments; a job asked to replace a timeline clip carries clip_landing: pending, then landed or not_landed with the reason. Unknown ids are omitted. Free.

