LabandFurnace.cc

Historical Mining

Back to dashboard

Historical Email Archive

Master Archive & Quote Mining

First sync Outlook messages into a deduped master archive, then classify and organize useful quote-learning conversations as candidates.

Supported files

PDF, Word, DOCX, Excel, XLSX

Included families

Muffle, vacuum, tube, rotary, kiln, drying oven, furnace variants

Excluded

Autoclave / sterilizer related emails

Email Archive Fetch

Builds one immutable master archive per inbox from Inbox, Sent Items, and Archive. Each Fetch continues from the newest archived email, so only missing newer messages are downloaded. Already archived messages are never re-downloaded or changed.

Fetch Start Date

Not set

Change in Admin Settings

Archive So Far

0 messages

Oldest: - · Newest: -

Fetch downloads only what the archive is missing, in both directions: emails newer than the newest archived message, and older emails back to the Fetch Start Date if the archive does not reach that far yet. Already archived emails are never re-downloaded or changed. If a fetch was interrupted, the next Fetch resumes it from its saved position automatically. The server keeps processing 500-message blocks after you click Fetch.

Email Archive Overview

No emails stored yet. Run Fetch to build the archive.

Candidate Classification

Classifies Sent Items into quote-learning candidates, starting from the Fetch Start Date configured in Admin Settings. Optionally set an end date to bound the run.

Starts From

Fetch Start Date not set

Change in Admin Settings

Soft Run

Test cleanup and model ratings on a small Sent Items sample. Each sampled email is rated separately by the configured OpenAI and Anthropic models so you can compare them side by side in Checked Email Review. This creates review rows and model ratings, but does not commit candidate seeds.

Current Process

Loading current process...