Four sub-topics were commissioned; two landed squarely on target, one produced a general evaluation methodology framework instead of the actual candidate data, and one researched an entirely unrelated product (kitchen CAD software branded “Winner Design”).
What the auth child established. YouTube’s cookie-based auth chain is now the only path for third-party tools — OAuth device-code login is deprecated for non-Google clients [1]. Since Chrome 127 (July 2024) shipped app-bound AES encryption, Firefox is the only browser whose cookies are reliably extractable by external tools [2]. More critically, Proof of Origin (PO) tokens are now mandatory for all YouTube video streams — a missing or invalid token returns HTTP 403, not a graceful auth error [3]. This means any archiver that does not run a current yt-dlp build with a PO token provider (e.g. bgutil-ytdlp-pot-provider [4]) will silently fail on video downloads regardless of how polished its subscription or UI features are.
What the Jellyfin child established. Channel-as-show output must follow Channel Name/Season XX/Channel Name - SXXEXX.mkv to be recognized as a TV series by Jellyfin’s scanner [5]. NFO sidecar files (tvshow.nfo at the channel root, per-episode .nfo) override all remote metadata and survive library rescans [6]. YouTube libraries must be type-segregated — mixing YouTube content with live TV or movies in one library degrades metadata matching [7].
The critical coupling. A YouTube archiver must solve both problems simultaneously: auth (PO tokens + Firefox cookies) and output naming (Jellyfin TV-show structure + NFO sidecars). Tools that handle auth but produce flat output — bare yt-dlp wrappers, basic download UIs — require a separate rename and NFO layer (ytdl-sub fills this role but adds configuration complexity). Tools with native Jellyfin output mapping must be verified to also keep their yt-dlp dependency current enough for PO token support. This two-axis coupling is where the candidate comparison should turn; neither a pure download UI nor a pure metadata layer satisfies both independently.
Watch Later is a partial-sync target at best. The YouTube Data API v3 permits playlistItems.list and playlistItems.insert for Watch Later (WL) but blocks update, delete, and reorder with 403 — enforced server-side [8]. Any archiver claiming “Watch Later sync” is doing authenticated yt-dlp polling of https://www.youtube.com/playlist?list=WL via browser cookies, not true two-way sync. Deletions from Watch Later in the YouTube UI cannot be mirrored to the local archive automatically. Rate limits compound this: authenticated sessions cap at ~2,000 videos/hour; unauthenticated sessions at ~300 [1].
Cookie rotation is the silent failure mode for subscription workflows. Exported cookies expire in approximately two weeks; leaving an authenticated browser tab open triggers background rotation scripts that invalidate already-exported cookies [1]. Any unattended archiver running channel subscriptions on a NAS needs a cookie-rotation strategy built in or bolted on — either scheduled Firefox cookie re-export automation or a dedicated throwaway Google account whose cookies can be refreshed without affecting the primary account.
The sharpest open question this expedition leaves unanswered: which candidates among TubeArchivist, Pinchflat, ytdl-sub, MeTube, YoutubeDL-Material, and Tartube handle PO token generation natively through their yt-dlp integration, produce Jellyfin-compatible TV-show output without a rename step, and have bus-factor > 1 with active releases in the last six months — all three simultaneously? Those three axes would determine the ranked pick; none of the four children answered them.