Epstein Studio

A collaborative workspace for annotating, discussing, and investigating redacted PDF documents. Built for the crowd-sourced effort to dig through the Epstein files.

Live at epstein-studio.com

---

What Is This

Epstein Studio renders PDF pages in the browser and lets users pin annotations directly onto the documents. Each annotation can include notes, styled text overlays, and arrow hints. Every annotation has its own discussion thread with voting, so the community can highlight what matters and surface the most promising findings.

Features

• PDF Viewer -- Pan, zoom, and navigate multi-page PDFs rendered as images
• Annotations -- Place anchor points on any page with notes, text overlays, and directional arrows
• Text Overlays -- Custom font, size, color, and opacity for marking up redacted or important sections
• Arrow Hints -- Draw arrows to point out connections or areas of interest
• Discussions -- Threaded comments on both annotations and entire PDFs
• Voting -- Upvote/downvote annotations and PDFs to surface the most promising leads
• Search & Browse -- Full-text search with autocomplete, sortable browse grid, random file picker
• Notifications -- Get notified when someone replies to your comments
• Heatmap -- Visual density overlay showing where annotations cluster on a page

---

Tech Stack

Layer	Tech
Backend	Django 5.2
Database	PostgreSQL 16
PDF Rendering	poppler-utils (pdftoppm)
OCR	Tesseract
Server	Gunicorn
Frontend	Vanilla JS, SVG canvas
Package Manager	uv
Deployment	Docker Compose

---

Getting Started

Prerequisites

• Python 3.9+
• PostgreSQL
• uv (Python package manager)
• System packages: poppler-utils, tesseract-ocr

Local Development

# install uv if you don't have it
curl -LsSf https://astral.sh/uv/install.sh | sh

# install dependencies
uv sync

# set up your environment
cp .env.example .env  # then edit with your DB credentials

# run migrations
uv run python backend/manage.py migrate

# index PDF files (point DATA_DIR to your PDF directory)
uv run python backend/manage.py index_pdfs

# start the dev server
uv run python backend/manage.py runserver

Docker

# bring up postgres + web server
docker compose up --build

# run migrations inside the container
docker compose exec web uv run python manage.py migrate

# index PDFs
docker compose exec web uv run python manage.py index_pdfs

Environment Variables

Variable	Description
DJANGO_SECRET_KEY	Django secret key
DB_NAME	PostgreSQL database name
DB_USER	PostgreSQL user
DB_PASSWORD	PostgreSQL password
DB_HOST	Database host (default: db in Docker, localhost for local)
DB_PORT	Database port (default: 5432)
ALLOWED_HOSTS	Comma-separated list of allowed hostnames
CSRF_TRUSTED_ORIGINS	Comma-separated list of trusted origins
DATA_DIR	Path to the directory containing PDF files

---

Project Structure

epstein/
├── backend/
│   ├── manage.py
│   ├── backend/              # Django project config
│   │   ├── settings.py
│   │   ├── urls.py
│   │   └── wsgi.py
│   ├── apps/
│   │   └── epstein_ui/       # Main application
│   │       ├── models.py     # Annotations, votes, comments, notifications
│   │       ├── views.py      # API endpoints and page views
│   │       ├── templates/    # HTML templates
│   │       └── static/       # JS, CSS, fonts, icons
│   └── email_header_extractor/
│       └── extract_headers.py  # OCR-based email header extraction utility
├── pyproject.toml
├── uv.lock
├── Dockerfile
└── docker-compose.yml

---

Contributing

This is an open investigation tool. If you want to help:

1. Fork the repo
2. Create a branch
3. Make your changes
4. Open a PR

---

License

MIT LICENSE