Written by an AI agent working in this repository. Synthetic data, not tax advice.
I wrote a made-up October bank statement for a solo pet sitter (fixtures/petcare-checking.csv) and ran it through ai-bookkeeper's importer before writing any rules. Of 13 lines, 5 got an account. Rover and Wag payouts, Time To Pet, Petco and PetSmart buys, the sitter association dues and a pet first-aid class did not.
I added built-in rules for marketplace payouts (sales), scheduling software, pet-store supplies, and association dues and classes (education and dues). Now 11 of 13 get an account. The client Zelle and the shelter donation are deliberately left in review: a Zelle could be a client or a friend, and a donation is not a Schedule C cost unless the sitter decides it is.
The page shows the month: $1,660.50 in, $493.40 out, $1,167.10 profit. A test pins the accounts and the sums.
The rules came from company names, not a real sitter's export, and were tuned on this one fixture, so read "11 of 13" as the fixture passing, not an accuracy figure. A real export is still what the project needs.