Email rules & tables

A lot of the email you retype by hand is the same shape, over and over: a Square receipt, a Shopify order, a class registration, a monthly invoice. An email rule teaches Dustav that shape once — who it's from and what to save — and from then on every matching email becomes a row in a table you can search, filter, and export. It's the data entry you keep meaning to automate, done by talking.

Make a rule by describing it

You don't build a template or wire up fields. You tell Dustav, in a sentence, what the email is and what to keep:

"I get a receipt from Anthropic every so often. Save the date, the receipt number, what it was for, and the total."

Dustav writes the rule — the sender/subject it should match, and the typed fields to pull — and shows you a real row read from a sample so you can check it before it goes live. Adjust it in the same conversation ("also grab the tax", "match only the ones about classes"), and it updates.

Typed fields

Each field has a type, so the data is usable, not just text:

• text — a name, a description
• number and money — quantities and amounts, stored as real numbers so they sort and total
• date — kept in a consistent format
• email, phone, url — the common structured bits

Picking the right type is the difference between a pile of text and a table you can actually work with.

One email, many rows

An email isn't always one record. An order with line items, a roster with ten names, a camp email listing every trip — Dustav pulls each row out of the one email and files them all.

The table

Every rule has a table on the Email tab: the rows it's collected, a search box, a status filter, and Export to CSV (the export follows whatever you've searched or filtered). Each row keeps a link back to the email it came from, so a number you're unsure about is one click from its source — nothing is a mystery cell.

A note on the table: it's read-only — you don't type into it. To fix or remove rows, just ask Dustav ("change the total on the Anthropic row to $18.40", "delete the duplicate receipts", "keep only the last ten"). It makes the edit and the table updates. If Dustav couldn't find a value in the email, it leaves that cell blank rather than guessing — a blank is a flag that something's missing, not a wrong answer dressed up as a right one.

Where the rows come from

A rule fills its table three ways:

1. Your daily Gmail read finds a matching email and files it (Connecting Gmail).
2. You forward a matching email to your Dustav address (Forwarding email).
3. You point Dustav at an email in chat and ask it to file the rows — handy for something already sitting in your inbox, or a dense PDF you want to walk through together.

What it is, and isn't (yet)

Today a rule gives you a typed table and a CSV export — a clean, exportable record of email you'd otherwise retype. It is not a dashboard: there are no charts or running totals inside Dustav yet. The types are captured now precisely so that richer views are cheap to add later, without re-reading a single email.