AI-Powered Epstein Network Analysis | An AutiTech LLC Public Access Project
Click Start Auto-Processing to begin analyzing DOJ documents, or use Analyze Document to paste text manually.
The network graph will populate as entities and relationships are extracted.
Loading network graph...
Building Epsteinpidia has been a labor of transparency and accountability. The construction of this project alone has already exceeded $500 in costs — money that has come directly out of my personal savings during a very difficult time. I haven't had steady work in a very long time, and this expense has eaten into my rent money. I am sincerely asking for your help to at least cover the construction costs so I can keep this project alive and continue making these documents accessible to the public. Every dollar truly makes a difference. Thank you for caring about truth and justice.
Processing...
Enter the administrator password to access protected records.
Incorrect password. Access denied.
Your donation helps maintain and improve Epsteinpidia, keeping this important research tool free and accessible to everyone.
Or enter a custom amount above
Select a preset amount for monthly recurring donations
Epsteinpidia is built and maintained by AutiTech LLC, a company that leverages the unique strengths of individuals on the autism spectrum to deliver innovative IT solutions. This project exists entirely thanks to community donations. Below is a transparent breakdown of the real operating costs.
AutiTech's mission is to showcase the value of neurodiversity in tech. Every donation directly supports keeping this important research tool free and accessible worldwide.
The DOJ Epstein Library contains 3.5 million pages of text documents, 180,000 images, and 2,000 videos across 12 data sets (~300+ GB on DOJ servers). Epsteinpidia processes all text content — approximately 2 million text-extractable pages. Extracted data (entities, relationships, locations, financial transactions, dossiers, evidence, and translations across 18 languages) is stored in a cloud PostgreSQL database (~25 GB at full capacity). The application runs on a Dedicated Reserved VM (2 vCPU / 8 GB RAM) for 24/7 pipeline processing and availability.
| Agent | Model | Daily Input Tokens | Daily Output Tokens | Cost/1M Input | Cost/1M Output | Daily Cost | Monthly Cost |
|---|---|---|---|---|---|---|---|
| Document Processing | gpt-5.2 | 12M | 4.5M | $1.75 | $14.00 | $84.00 | $2,520 |
| Dossier Agent | gpt-5.2 | 300K | 100K | $1.75 | $14.00 | $1.93 | $58 |
| Translation (17 Agents) | gpt-5-mini | 2.55M | 2.55M | $0.25 | $2.00 | $5.74 | $172 |
| AI Processing Subtotal | $91.67 | $2,750 | |||||
| Category | Monthly | Yearly | 5-Year |
|---|---|---|---|
| Replit Core Plan | $25.00 | $300 | $1,500 |
| Dedicated Reserved VM (2 vCPU / 8 GB RAM) | $80.00 | $960 | $4,800 |
| PostgreSQL Storage (25 GB @ $1.50/GB) | $37.50 | $450 | $2,250 |
| PostgreSQL Compute (24/7 pipeline — 720 hrs @ $0.16/hr) | $115.20 | $1,382 | $6,912 |
| Data Transfer (~150 GB @ $0.10/GB) | $15.00 | $180 | $900 |
| Domain Registration | $1.00 | $12 | $60 |
| Stripe Processing Fees | ~2.9% + $0.30 per transaction | ||
| Infrastructure Subtotal (excl. Stripe) | $273.70 | $3,284.40 | $16,422 |
| Data Category | Estimated Size | Details |
|---|---|---|
| Document Notes & Summaries | ~1.2 GB | ~600K documents × ~2 KB each |
| Translations (17 Languages) | ~21 GB | All translatable content × 17 language agents |
| Entities & Relationships | ~350 MB | ~100K entities + ~500K relationships |
| Token Usage Records | ~1 GB | ~5M API call records |
| Dossiers & Web Citations | ~300 MB | ~10K dossiers + ~200K citations |
| Locations, Financials, Other | ~150 MB | Activities, transactions, donations |
| Total Database Storage | ~25 GB | $1.50/GB/month = $37.50/mo |
Powered by AutiTech LLC — showcasing the value of neurodiversity in technology.