955.8K
HEWhat looks like “gibberish” in the Epstein files is Base64, a format used in MIME emails to embed attachments like PDFs and images.
Digital forensics enthusiasts have already shown some of this data is recoverable. One of them, Mahmoud Al-Qudsi (mqudsi on X), has been documenting the process in real time.
In short:
• Base64 strings were embedded in DOJ PDFs (that probably shouldn’t have survived redaction)
• OCR failed because it tried to make words — Base64 is not human readable
• Even tiny errors (like 1 vs l in Courier New) break decoding
• A CNN trained on pixel-level characters solved most issues
• Comparing duplicate email threads fixed the rest
The attachment successfully decoded was a PDF. The contents weren’t that important, but the process matters.
It shows:
– some attachments are recoverable
– redaction/export workflows were sloppy
– more Base64-encoded files are likely still reconstructable
Mahmoud is sharing the full technical breakdown on X if you (understandably) don’t want to support X, he also has a Substack.
This is slow, careful work. Please don’t spread viral claims without fact-checking, misinformation only creates noise and shields people who deserve real accountability.
More updates to come.
@helloitsmia.tech
