RYE HOWARD-STONE
Document forensics for investigations too big to read by hand.
I run your corpus — FOIA dumps, court records, leaked archives, public-records scrapes — and return a sourced findings memo with stable document IDs, reproducible queries, and alteration forensics. You keep the byline. You verify everything.
$ casestack scan --case doj-epstein-2025 --check gaps
comparing document identifier sequence across federal releases…
FINDING ~64,000 identifiers present in the Release 1 index
are absent from Release 2. Removed without notice.
SCOPE 1,380,000 documents / ~3,000,000 pages — full read, not sampled.
METHOD deterministic identifier-sequence diff. Reproducible:
casestack scan --check gaps --diff release1 release2§ 01 Who this is for
- Solo investigative journalists and 2–8 person newsrooms sitting on a document set too large to read before deadline.
- Independent and Substack investigators competing against better-resourced outlets.
- Anyone holding a FOIA release, court archive, or leak that contains a story they can't physically extract alone.
§ 02 What you get
- A sourced findings memo — every claim tied to a stable document ID, a reproducible query, and a page reference.
- A searchable full-text database of your corpus — the same engine behind epstein-data.com, pointed at your documents. Built on CaseStack, open source.
- Alteration & gap forensics where applicable — what was removed, changed, or is missing, with a methodology you can print.
§ 03 How it works
- 01 Send me the corpus — or where it lives — and the question you're chasing.
- 02 I run it and return the findings memo with stable citations.
- 03 You verify, you publish, you keep the byline. The methodology stays open.
§ 04 Engagements
Search Build
$400 flat
Your document set — up to a few thousand pages — ingested, deduplicated, OCR'd, and handed back as a searchable database, with a 20-minute orientation. The fast one.
Buy nowFindings Memo
from $1,500 scoped on the call
The full forensic pass: entity linking, deduplication, gap and alteration analysis, and a sourced memo where every claim is citable and reproducible. Journalism-grade output.
Book the callLarge corpus / commercial
Custom quoted after scoping
Hundred-gigabyte-class releases, ongoing engagements, or running CaseStack commercially in-house. We scope it together on the call.
Book the call§ 05 Who's doing the work
I built epstein-data.com — 1.38 million documents, roughly 3 million pages, full-text searchable, used by reporters at WIRED and NPR. I found about 64,000 documents silently removed from a federal release, because I could read the whole corpus and a person can't. I hold a PhD in computer science and have spent a decade on large-scale document and sequence analysis. Everything here is done by me, not outsourced.
§ 06 Questions
- How long does it take?
- Scoped on the call — it depends on corpus size and the question. Small sets turn around in days.
- What do you need from me?
- The documents (or portal access) and the question you're chasing.
- Can my editor trust the output?
- Every claim carries a stable document ID and a reproducible query. You verify before you print; the methodology is open.
- What if my document set is huge, weird, or encrypted?
- That's the normal case. Book the call and we'll scope it.
If you're sitting on documents you can't read in time, the story doesn't wait.
Book a free 30-minute call