nx_archive_research.nx

nx_archive_research.nx -- THIN structured TIME-MACHINE / WEB-ARCHIVE research source organ. COMPOSES nx_research_engine (sovereign fetch->extract->bank). Operator 2026-06-29: "browse all the historic page3.com or megastar.co.uk sites day by day as they existed back then with their media ... the wayback machine has some capabilities but some media archival was hit or miss ... same for historic reddit.com." GOAL: a sovereign archive that EXCEEDS the Wayback Machine on (1) media completeness via cross-archive + perceptual recovery, (2) temporal granularity (day-by-day), (3) rendering fidelity, (4) provenance / no-delete. Ground the build on REAL sourced practice -- WARC / Common Crawl, Wayback / Memento, PageRank link-analysis, perceptual-hash reverse-image (TinEye), content-addressed storage -- never assumptions (Rule 4). Banks clean .txt -> knowledge/library/ (nx_library_harvest_v2 indexes, nx_search_cli queries). Canonical fact-dense static pages over sovereign TLS-1.3 fetch. Prefix arc_ keeps this distinct from the existing srch_ corpus. expect_exit: 0 license_tier: ORIGINAL