CAINC
See the shape of what an institution says
CAINC turns a large collection of institutional documents into a Semantic Terrain: a map you can explore. Topics become peaks, the boundaries between them become real edges, and years of priorities become visible at a glance.
What is a Semantic Terrain
A corpus has a shape. CAINC reads it.
Every large document corpus contains structure. Topics cluster. Some dominate; others are peripheral. Enforcement priorities shift over time. Narratives converge and diverge.
A Semantic Terrain is the computed reading produced by CAINC. You navigate it the way you navigate geography: by moving through a landscape with real topographic features, not by entering keywords into a search box.
Each terrain is built from the full corpus of a named institution. It is updated as the corpus grows. It is the same document body, seen differently.
The dominant topics of the corpus. Each peak is a stable cluster of language that defines a region of institutional focus. Drill into any peak to reveal finer structure within it.
The boundaries between topics. Where one enforcement focus ends and another begins. Computed from the actual geometry of the corpus.
The transition zones between peaks, where topics overlap, share language, or shift into one another.
The paths connecting related peaks. Follow a corridor to see how two enforcement areas share language, entities, or structural proximity.
Groups of related peaks that form broader regions of institutional focus. Communities are organized in tiers — from broad continental themes down to fine-grained street-level neighborhoods. Each community is named for its dominant content and navigable as a scope.
People, organizations, and cases located precisely within the terrain. Find any entity and see its full dossier: co-occurrences, peak membership, and related documents.
Live · Department of Justice
DOJ Corpus Terrain
The full body of DOJ public communication, geometrically structured and navigable. See which enforcement topics dominate by era. Locate any defendant, case, or charge type within the broader terrain of federal prosecution. The DOJ Terrain is our public reference reading, open to anyone.
Who uses Semantic Terrains
Journalism & Investigation
Journalists and Investigators
See years of enforcement priority without reading years of press releases. The shape of the corpus tells you what an institution cared about, when, and how much.
- →Spot enforcement priority shifts across administrations
- →Find the topographic neighborhood of any charge type or case
- →Surface structural patterns invisible to keyword search
Legal & Litigation
White-Collar Defense Attorneys
Map how the DOJ frames similar conduct across time and jurisdiction. See where your case sits within the full landscape of federal enforcement and what surrounds it.
- →Locate structurally similar prosecution narratives
- →Identify proximity between charge types in the terrain
- →Track how enforcement language around a topic has evolved
Compliance & Risk
Compliance Analysts and Risk Officers
Detect emerging enforcement focus before it becomes a crisis. Monitor how regulatory language around your sector is moving relative to the full corpus and where it is heading.
- →Identify rising peaks in enforcement terrain
- →Track topic boundary shifts as regulatory focus moves
- →Monitor entity proximity to major prosecution clusters
What you can do in CAINC
Four ways to explore a Semantic Terrain
Terrain Map
See the full topographic landscape of the corpus. Peaks, contours, and communities rendered as navigable geography. Zoom from broad institutional themes down through regional patterns into fine-grained topic neighborhoods. Click any peak to explore its documents. Follow corridors between peaks to trace how topics connect.
Network View
Explore how entities relate to each other and to the terrain. Entities are positioned according to their location in the landscape, not arranged arbitrarily. See co-occurrence connections, community membership, and the structural neighborhood around any person, organization, or case.
Article View
Search the full corpus by meaning, or search within any scope you're exploring — a peak, a community, an entity's documents. Results are ranked by semantic similarity, surfacing structurally related documents even when they use different language.
Entity View
Pull up any entity in the corpus and see its full dossier: every document it appears in, which peaks it belongs to, which other entities it co-occurs with, and how its presence in the terrain has developed. Dossiers are scoped — open an entity from within a peak or community to see its role in that context.
How Semantic Terrains work
Computed from the documents, not drawn for effect.
Many document maps are drawn for effect: the contours look meaningful, but they come from the rendering, not from the documents.
CAINC computes the landscape from the documents themselves, then finds its peaks, passes, and boundaries. What you see reflects how the language actually clusters.
The mathematical substrate is documented in full on the Methodology page. For definitions of terrain features, see Terminology.
Documents are ingested from the source institution
The full document corpus of a named institution is collected, cleaned, and prepared for analysis. Every document is preserved in full.
Documents are placed by meaning
Similar language lands together. Distinct topics separate. Every document keeps its place as evidence.
Topics, entities, and boundaries are identified
Peaks, passes, contours, corridors, and the skeleton that connects them are computed from the actual geometry of the corpus. Communities are detected at multiple scales. Entities are located precisely within the terrain.
The terrain updates as new documents are published
As the live corpus grows, the Semantic Terrain is recomputed on its defined cadence. You see terrain movement, not static snapshots.
You navigate by topic, entity, time, and proximity
Explore the terrain by moving through its geography. Find any entity and see what surrounds it. Follow a topic across time. Trace the boundary between two enforcement areas.
Available Terrains
Each terrain is a named corpus. One institution. Full depth.
Terrain Subscriptions give you continuous access to a specific institutional Semantic Terrain built from a named corpus, updated as new documents are published, and navigable by entity, topic, and time. The DOJ Terrain is our permanent free reference reading. Named terrain subscriptions are in development.
Department of Justice press releases, 2020 to present. 116,000+ documents. The full landscape of federal prosecution, enforcement priority, and institutional communication.
Drug Enforcement Administration. The full terrain of federal narcotics enforcement: charge types, priority shifts, geographic and entity patterns across decades.
Federal Trade Commission. Consumer protection and antitrust enforcement terrain for compliance, M&A due diligence, and competitive landscape research.
Supreme Court of the United States. Published opinions and argumentation. The structural shape of constitutional doctrine across two centuries of language.
Terrain Subscriptions provide continuous access to a specific institutional Semantic Terrain built from a named corpus, updated on a defined cadence as new documents are published. Access details are handled by direct inquiry ahead of each terrain launch.
Custom Terrains and Deep Dives are available for organizations that need a specific body of documents mapped or a scoped analytical engagement. These are handled by direct inquiry.
Interested in a specific terrain?
Tell us about your corpus, your questions, or the institution you need mapped.