Mixed per-item model: each audio/lyric/story entry can be marked as (a) contributor‑owned with permission, (b) public domain, or (c) requires permission. We will add clear per-item credits and takedown workflow.