Canonical queued-ingest endpoint (Phase 2B, 2026-05-25).

<br/>Body: { url: "https://instagram.com/reel/..." }<br/><br/>Creates a CopyHistory row in 'queued' state, hands off to the<br/>run_queued_ingest Celery task, and returns 202 + the ingest_id<br/>so the frontend can poll GET /api/v1/copy-ingest/history.<br/><br/>Dedupes per (owner, url): re-posting the same URL from the same<br/>user returns the existing row's id with its current status (200,<br/>not 202) so the client's poll loop keeps working.<br/><br/>Worker handoff: the row stays in 'queued' state until the Celery<br/>task picks it up (app/modules/copy_ingest/tasks.py::run_queued_ingest).<br/><br/>level picks the extraction tier (transcript|scene|full|replica).<br/>'replica' (Clone-a-Reel, 2026-07-10) adds the per-shot source<br/>breakdown used to auto-build a replicated storyboard; re-posting a<br/>URL with a higher level than its existing row upgrades that row<br/>in place.<br/><br/>429 COPY_INGEST_DAILY_CAP once the caller has started<br/>COPY_INGEST_DAILY_CAP new copies inside the window (20 by default,<br/>admins exempt). A dedupe and a free clone of a ready analysis never<br/>count; see daily_cap.py.