You know the file exists. You know roughly who sent it. What you don't know is which of the nine emails has the current version. This indexes every attachment in your mail and groups them by document, newest first.
Worked example: a scope of work that's been round the houses — you have
Scope_of_Work.docx, Scope of Work v2.docx,
scope-of-work-FINAL.docx and Scope of Work final (1).docx,
spread across three threads and two senders.
It normalises the file name and strips the noise people add to versions:
| Stripped | Examples |
|---|---|
| Version tokens | v2, rev, final, draft |
| Copy counters | (1), (2) |
| Embedded dates | 2026-03-14, 031426 |
| Separator differences | Scope_of_Work = Scope-of-Work = Scope of Work |
That last row matters more than it looks. Separators are normalised to spaces
first — before anything else — because a word boundary doesn't fire next to an
underscore, so Scope_of_Work_v2 would otherwise keep its v2 and
split off as its own document. It's the kind of bug that produces a tidy-looking list that
is quietly wrong.
Scope of Work became
SOW - District 4. Nothing can group those by name alone. Search by sender
instead and you'll find both.Mail.Read). It cannot send, delete, move, or modify anything — not your mail,
not your files, not the attachments themselves. It builds an index and shows it to you.
If it could do more, it would need permissions it deliberately never asks for.Attachment contents are never downloaded while indexing — only the file names, senders and dates it needs to group things. There's no publisher server and no telemetry.