easterbrooka teaches Mike to read the attachments, not just the email
Now when the AI opens a .msg or .eml, it sees what's stapled to it too.
Anyone who's used a legal-AI tool on an email thread knows the dance: forward the message, then dig out every PDF and Word doc and upload them as separate files before you can ask a single question. easterbrooka's fork cuts that step out. When the model opens an email, it now also pulls in the readable text of the attachments - contracts, spreadsheets, even nested emails - and hands the lot to the AI in one go.
There are sensible limits: the fork skips inline images so signature logos don't muddy the prompt, caps how much text any one attachment or email can contribute, and only goes three layers deep into forwarded-thread chains. Nothing changes in the interface - same email view, same attachment chips. The difference shows up only when you ask a question.
Spotted something wrong? Or know the PR text has fresher detail than the writeup above?