Account Transactions

This data model extracts individual transaction details typically found on bank or credit card statements. It captures key information for each transaction, including the date it occurred, the description provided on the statement, the monetary amount (standardized so that money coming in is positive and money going out is negative), and categorizes the transaction into both a specific type (like groceries or wages) and a broader group (like food_and_drink or income) based on a standard list.

Field name	Type	Description
transactions	array	A list containing details for each transaction found on the statement.
transaction_date	string	Date the transaction happened or was officially recorded by the bank, shown as YYYY-MM-DD.
description	string	The exact text describing the transaction as it appears on the statement.
amount	number	The monetary value of the transaction. Positive numbers indicate money coming into the account (like deposits), while negative numbers show money leaving (like payments or withdrawals).
category	string	A specific category assigned to the transaction based on its description, chosen from a predefined list (e.g., `groceries`, `wages`, `car_payment`).
primary_category	string	A broader category assigned to the transaction based on its description, chosen from a predefined list (e.g., `food_and_drink`, `income`, `loan_payments`).

Field Notes

transaction_date: Must be in YYYY-MM-DD format.
amount: Extracted numeric value. Positive for credits/inflows, negative for debits/outflows. Extracted to a maximum precision of 2 decimal places.
category & primary_category: These fields are assigned based on the transaction description using the specific taxonomy provided in the extraction prompt. Both values must be one of the exact snake_case category names listed in the taxonomy. If a suitable category cannot be determined, these fields will be null.
- Allowed primary_category values: trade, income, transfer_in, transfer_out, loan_payments, fees_commissions, entertainment, food_and_drink, general_merchandise, home_improvement, medical, personal_care, general_services, government_and_non_profit, transportation, travel, rent_and_utilities.
- Allowed category values are the specific sub-items listed under each primary category in the extraction prompt's taxonomy (e.g., buy, sell, dividends, interest_earned, groceries, restaurant, car_payment, atm_fees, etc.).