Multi-Agency Document Intelligence — Built, Tested, Validated.
One search. Every agency. Every document. AI-powered analysis with zero data leakage.
DeclassDB is the public proof. The same architecture — cross-agency search, full-text PDF search, entity extraction, AI analysis, and local-first privacy — scales from declassified public records to any controlled-access, cross-security (Confidential → Secret → Top Secret), cross-agency document environment.
Validated on the hardest public dataset available
Real declassified records across seven federal agencies — these are measured numbers, not projections.
What the platform does
Cross-Agency Search
Unified search across any agency document system — FBI, DEA, ATF, CBP, DOJ, State, DoD. One query, every source, merged and de-duplicated.
Full-Text PDF Search
OCR + full-text indexing across 1.05M CREST pages. Search inside documents, not just metadata — down to the page.
Entity Extraction & Graph
Automatic entity identification — people, organizations, locations, dates, case numbers — with cross-document relationship mapping.
AI Analysis — Localized & Containerized
Summarization, pattern detection, and investigative leads from on-prem, containerized AI models running inside your boundary — no cloud dependency, no data egress, air-gap-capable. (BYOK cloud optional for unclassified tiers only.)
Workflows & Compliance
Request generation, audit trails, and citation export (BibTeX / RIS). Compliance-ready from day one.
Local-First / Zero Trust
Browser-local storage. No server-side user tracking, no search logs. Runs air-gapped, on-prem, or sovereign cloud — DDIL-ready.
Where it applies
- Cross-border & multi-agency investigations
- Task-force coordination across jurisdictions
- Unified case files from disparate systems
- Evidence chain & document provenance
- Strategic intelligence aggregation
- Declassification & FOIA review at scale
- Cross-agency briefing preparation
- DDIL / disconnected field operations
- Regulatory audit & e-discovery
- Controlled-substance / financial monitoring
- Internal investigations
- Multi-jurisdiction document management
Cross-security by design
The classification of a document is metadata, not architecture. The same engine that searches public declassified records enforces compartmentation across Confidential, Secret, and Top Secret tiers — access-gated per user, per agency, per caveat — while keeping the data local-first and auditable. Cross-functional, cross-security, cross-agency: one platform, every clearance level.
DeclassDB.com is a free, public instance operated by ResilientMind AI LLC (SDVOSB). The classified / controlled-access deployment is delivered as the AriaOS Classified System via CommandRoomAI.com.
DeclassDB
FOIA search is fragmented and keyword-only. DeclassDB unifies seven federal archives — CIA CREST, FBI Vault, NSA, State, NARA, DoD, NSArchive — and adds AI semantic search, so you can find what you mean, not just what you typed.
AI semantic search
Find by meaning across 309,708 CREST documents (1.05M pages, full-text). The embedding model runs in your browser — no cloud round-trip. Pro & Researcher.
One query, every agency
Unified, blended, de-duplicated across seven federal sources. CIA / FBI / State always free; NSA / NARA / DoD / NSArchive unlocked on Pro.
Privacy-first by design
Vectors, AI summaries, and entity extraction all run on your device. No cloud, no search logs, no third-party trackers. Air-gappable for federal use.