The whole database, in plain text

Parcha's canonical store is a set of human-readable YAML files in a public Git repository. Every build compiles them into SQLite and dumps the result as plain SQL — so the entire database, including its history, survives as text. If this site disappears, the record doesn't.

parcha.sql

Full SQL dump

Schema + all 80 incidents, sources, legal actions, actors, status logs. Loads into any SQLite.

parcha.json

Site JSON

The denormalised feed this site renders from — one object per incident, generated 2026-06-11.

Reproduce it

git clone https://github.com/cookedzera/Paperleak
cd Paperleak
python3 cli/parcha.py validate   # run the publication gates
python3 cli/parcha.py build      # YAML -> SQLite + SQL dump + JSON

Licence

Data: CC BY 4.0 — use it for anything, credit "Parcha". Code: MIT. If you publish work built on this data, we'd love a link, and we'll list it.

Citing a single fact

Every case page cites a dated document per row. Cite the case URL plus the document on that row — not this site alone. The register is a map to primary sources, not a substitute for them.