Skip to content

Attachments and documents

There is one file store, reached two ways: from the record a file belongs to, or from the library that lists them all.

Attachments

An attachment belongs to a record. You add it from that record's page — a receipt on an expense, a signed contract on a deal, a scan on a contact — and it stays with it.

Every attachment carries, beyond the file itself:

  • title and description, for when scan_003.pdf is not a name;
  • document type, your own classification;
  • document date — the date on the document, which is rarely the date it was uploaded and is what you will search by;
  • tags, the same tags used elsewhere;
  • party, the other side of the document — who issued it, who it is for;
  • keep until, a retention date, for the documents you are obliged to hold for a fixed number of years.

All of it is optional. A file uploaded and never described is still a file attached to the right record, which is most of the value.

The Documents library

Documents lists every previewable attachment you can see — PDFs and images — across every record type at once, each row showing which record it hangs off.

The restriction to previewable types is what makes the library coherent: it is for things you would look at, not a file manager. A spreadsheet attached to a deal is still there on the deal; it simply is not part of a library you browse visually.

Because the library spans record types, its filters are about the document rather than its owner: document type, date, tags, and which kind of record it is attached to. It also counts what is untyped, which is the useful list — the files you uploaded in a hurry and never classified.

Which way in?

Go through the record when you know what the file relates to. Go through the library when you know something about the file — roughly when, roughly what kind — but not what it was filed against. The second case is the one a per-record view cannot answer, and the reason the library exists.

A note on counts

The library's "more results" indicator is a hint rather than a promise: the permission filter runs after the query, so an exact count would mean fetching everything you cannot see in order to not show it. Paging works; the total may shift by a row.

Kancil — one workspace for contacts, work and money.