Just set a cron job up to scrape and md5 all of the documents every 20m or so; if the MD5s match, discard. If they do not, save. Then you can build a timeline of releases, redactions, edits, and fuckery.

UPDATE: As of 2:17pm ET on 1/30/26, the document is once again appearing on the DOJ website. We are looking through to see if additional redactions were made.

“Oops.”

Vile.

There’s this one too

https://www.justice.gov/epstein/files/DataSet 10/EFTA01660651.pdf