Millions of pages of evidence are public.
The investigation starts here.
AI-assisted analysis of the full Epstein public archive. Not search results — structured, confidence-rated, legally-mapped findings.
Individual A appears in flight logs alongside Individual B on 14 documented occasions between 1999–2005
- 2024 FOIA flight log release
- SDNY filing 2019
- Maxwell trial exhibit 44
Corroborated across 3 independent sources. No inference required.
Assessed against 18 U.S.C. § 1591. Elements present: 3 of 5.
Illustrative output only. No real findings published without meeting defined evidentiary threshold and passing independent legal review.
The archive spans 25 years, multiple countries, and dozens of jurisdictions. No single institution has reach over all of it.
The volume defeats human analysis. Millions of pages were released without context, without descriptions, without anyone whose job it is to connect them.
That is not an accident of history. It is a structural gap. And it is exactly what AI now makes possible to close.
"The Epstein Files are public. Now we do the work the system refused to do."
What Exists — and What We're Building
Serious people have built serious tools. Search engines, vector databases, email explorers, document indexes — they let anyone find what they already know to look for. That work matters. It made this next step possible.
But retrieval is not reasoning. Finding a document is not the same as understanding what it means in the context of ten thousand other documents across twenty-five years. No one has built the layer that connects evidence into findings — that cross-references individuals, timelines, and legal frameworks across the full archive.
We are not replacing what exists. We are building on top of it. This is how science works. This is how investigations work. Each generation of tools makes the next one possible. The search engines proved the archive could be navigated. We are building the system that reads it.
We do not know what the findings will show. That uncertainty is not a weakness — it is the definition of honest investigation. If the evidence is inconclusive, that itself is a meaningful result for the victims and for public understanding.
What The System Produces
Raw evidence moves through three specialised layers of reconstruction — then converges into structured, source-cited, legally-mapped findings.
Timeline Reconstruction
A reconciled timeline of every documented event — sequenced, cross-referenced, with gaps identified and flagged.
Relationship Mapping
A documented map of every connection between individuals in the archive — who appears with whom, how often, across which documents and years.
Legal Framework Mapping
What the evidence supports — mapped against specific statutes, scored element by element, published with full source citation.
STRUCTURED FINDING
ILLUSTRATIVE EXAMPLE — NOT A REAL FINDING
Illustrative data only. All persons, entities, dates, and document references shown above are fictional examples. No real findings have been published. Nothing will be published without meeting a defined evidentiary threshold and passing independent legal review.
What Your Support Makes Possible
This is a multi-year project. The archive is not static — new documents are released continuously through ongoing litigation, and a credible investigation will itself trigger further releases requiring retroactive analysis. Doing this fully may cost millions considering the volume of work and the cost of AI. We are starting small — as any serious initiative would — to prove the system works, then raise more.
It funds the infrastructure build — ingestion pipelines, OCR reconciliation, entity extraction. It pays for the first analysis sprint across the highest-priority document sets. And it establishes the legal structure that governs how findings are reviewed, verified, and published.
If the goal is not reached, contributions are returned. No ambiguity.