history_service

Unified historical archive search. In-memory loading of 4 sources with fuzzy address matching.

Connections

• sources: Anomaly Archive (42 GB), Blackhouse Archive (17.5 GB), Dropbox Archive (19 GB), Xero Archive (12 MB)
• exposes: 10 /history/* intelligence endpoints
• pure Python: no external libraries

Notes

• 10 tools: precedent, pricing, repeat-check, recon, morning-context, production, permit-timeline, communications, invoice-search, jurisdiction
• Fuzzy address matching via difflib
• ~85 GB across 5 archive sources, matched to Mostly projects
• In-memory for speed — loads all archives at startup