The 28 data sources and agents feeding the Hub database (mjsgtszehjltxmbxtctz), grouped into 5 swimlanes: life science domain APIs, news and market signals, content and web-search learners, deep research plus the RAG/synthesis layer, and BioCreative's own operational signals (SEO/GEO, meetings, email, calendar). These are parallel, independent sources, not a sequential pipeline. Several feed directly into the Account & Contact Lifecycle diagram's Phase D signal loaders. Verified live against both VPS crontabs (bc-ops, bc-kb) and n8n on 2026-07-26 — not just skill-file claims.
7 systems, all free public APIs, forming the life-science backbone that also feeds account enrichment and the outreach signal loaders
Cross-feed correlation and news monitoring, synced to client spokes on a per-client cron schedule
Extract, classify, and catalog pattern applied to video, social, code, Reddit, RSS, and the open web
Where everything above becomes queryable strategic intelligence, or gets synthesized into daily/weekly content proposals
Not life-science data — BC's own search performance, meetings, email, and calendar, all live-cron verified on bc-ops and bc-kb