v1 — Released February 25, 2026. The Jmail Data API is stable and ready for production use. See the changelog for details.
data.jmail.world — no API keys, no rate limits, no authentication. Visit data.jmail.world for manual access to the data.
What’s Available
| Dataset | Description | Records | ~Size |
|---|---|---|---|
| Emails | Full email archive with body text, sender, recipients, dates | 1.78M | 334 MB |
| Emails (slim) | Network-only view — no body text, much smaller | 1.78M | 41 MB |
| Documents | Document metadata (filenames, descriptions, page counts) | 1.41M | 25 MB |
| Documents (full text) | Sharded full extracted text from all documents | 1.41M | Large |
| Photos | Photo metadata with AI-generated descriptions | 18K | ~1 MB |
| People | People identified via facial recognition | 473 | <100 KB |
| Photo Faces | Face bounding boxes linking photos to people | 975 | <100 KB |
| Star Counts | Crowd-sourced star/interest counts | 414K | ~2 MB |
| Release Batches | Release batch metadata (names, dates) | — | <10 KB |
How to Access
Python Client
uv run.DuckDB / SQL
Query Parquet files directly over HTTP with SQL. Zero download.
Direct URLs
Download Parquet or NDJSON files directly.
The Jmail Ecosystem
The Data API gives you programmatic access to the same data powering the full Jmail suite:Jmail
JDrive
JPhotos
JFlights
JMessages
Data Sources
All data in Jmail comes from three primary sources:- House Oversight Committee (November 2025) — Gmail emails, converted from PDFs to structured text
- Department of Justice (December 2025 – January 2026) — Documents, photos, and emails released per the Epstein Files Transparency Act
- DDoSecrets — Yahoo emails published in collaboration with Drop Site News, added in weekly batches
Base URL
All datasets are served from:Formats
Every dataset is available in two formats:- Parquet (
.parquet) — columnar binary format, best for analytics - NDJSON (
.ndjson.gz) — gzipped newline-delimited JSON, for streaming/scripting