About CRIMENET

CRIMENET maps criminal organizations and their relationships into a single, searchable graph. It covers cartels, mafias, gangs, motorcycle clubs, triads, clans, factions, militias, and terrorist groups, drawn from 1,418 Wikipedia articles in English, Italian, Portuguese, and Spanish.

An LLM pipeline reads each article and extracts every organization mentioned and every alliance, rivalry, or other tie between them. Every edge carries the verbatim Wikipedia sentence that justifies it, a paraphrased description, an extracted time period, and a versioned source URL. Any claim can be audited back to the specific revision of the specific article it came from.

There is, to my knowledge, no larger directory of criminal organizations anywhere. Wikipedia's own list of criminal enterprises, gangs, and syndicates covers a few hundred groups. A Google search for "all criminal organizations worldwide" returns the candid admission that "an exhaustive list of all criminal organizations globally does not exist." CRIMENET comes closer than anything else: nearly 5,000 organizations mapped across nearly 11,000 relationships, each one backed by a specific Wikipedia source.

This was an accidental achievement. The goal was to build a Knowledge Graph of how criminal organizations relate to each other, not to catalog every group mentioned on Wikipedia. But because the LLM pipeline reads 1,418 articles across four languages and extracts every organization mentioned in each one, it ended up capturing the vast majority of criminal organizations documented on English, Italian, Portuguese, and Spanish Wikipedia. CRIMENET became the comprehensive list that did not exist before.

The graph reflects what was found in the articles processed. Not every organization retrieved in the process is necessarily a criminal organization. Some state forces, political parties, or other non-criminal entities may have slipped in. The LLM pipeline is not perfect and occasionally misses a connection or misclassifies one. More importantly, Wikipedia itself does not document every real world tie. An absent link does not mean a relationship does not exist; it means it was not recorded in the sources that were processed. And while CRIMENET is likely the most extensive catalog available, it is certainly not exhaustive: organized crime is by nature hidden, and many groups operate below the threshold of public documentation.

The code, the dataset, the pipeline, and a detailed technical report are all open source. The GitHub repository has the full pipeline, the audit tools, and the build system that turns the data into this site.

When to use what. For specific, structured queries (looking up a single organization, finding all edges between two specific orgs, seeing which orgs operate in a given country, browsing community clusters, or exploring triadic signals), use the dedicated tabs on the connection finder and the dashboard. These tabs render the data directly with no LLM involvement. For complex questions that combine multiple pieces of information (tracing indirect relationships through intermediaries, comparing organizations across countries, ranking by network influence, or asking about structural patterns), use CRIMENET AI. The AI queries the graph's structure through tools and synthesizes an answer. As with any LLM response, verify claims against the source evidence the answer cites.

This is a complete example of GraphRAG, retrieval-augmented generation over a Knowledge Graph. Instead of searching documents by keyword similarity, the AI queries the graph's actual structure: it looks up organizations, traces connections through intermediate nodes, finds shortest paths, inspects communities, and ranks by network centrality. This lets it answer questions that flat document search cannot: questions about indirect relationships, network position, community membership, and structural patterns that only emerge from the graph. You can ask questions like "Which Mexican cartels operate in Europe?" or "How are the PCC and 'Ndrangheta connected?" and get evidence backed answers drawn directly from the graph. The AI calls tools that search organizations, trace connections, find paths, and inspect relationship summaries, all running in the browser against the same static data files that power the rest of the site. Every answer cites the underlying Wikipedia sources, so claims remain traceable and auditable.

An interactive world map of cross-border organized crime. Each country shows the organizations based there and foreign groups operating within its borders, drawn from CRIMENET's documented country footprints.

An interactive 3D network of the full graph, nearly 5,000 organizations and nearly 11,000 relationships rendered as a force-directed layout. Pan, zoom, and hover to explore how criminal organizations connect.