Decision layer above the live stream. Priority briefs and executive summary load from the same cached narratives as the intel story rail, re-ranked with live market context.
Impact and confidence scores, signal decay hint, why-it-matters lines, rank factors, and tickers.
Story threads carry status, update counts, first/last seen, and impact/confidence trends. Clustering uses tickers + headline Jaccard; timeline lists first mention vs follow-ups.
Locked hub days (UTC), oldest first — this is hub archive index, not a live activity log. Open a day in the Archive tab for the full snapshot.
Your working memory — clip news, ideas & risks. Drop files, let AI brief them, organize by board.
Scrapbook audit trail: saves, board moves, note edits, and deletes (separate from History, which lists locked hub archive days).
These settings are stored in this browser only (localStorage). They tune the hub API query string: watchlist symbols, how many briefs/stories to rank, item pool size, and thread clustering sensitivity.
Lock a daily snapshot of priority briefs, executive summary stub, and story threads (same pipeline as the live hub, no extra LLM). Each calendar day is stored once per org/user. Weekly/monthly rollups can layer on this table later.