v1

latestOpenAPI 3.1.02026-07-26232644.9 KB
Transactions

Retrieves the user's financial transaction history within a specific period with intelligent filtering and automatic categorization. Supports filtering by account type, transaction type, category, amount ranges, and natural language queries via clientMessage parameter. Applies two processing stages: 1) LLM-powered intelligent filtering (when clientMessage is provided), 2) Automatic transaction categorization (always applied). Supports two response formats: 'raw' (default) for unprocessed data, and 'structured' for organized data with groupings and summaries. Requires API key for authentication and active subscription.

get/tools/api/get-transactions

Query parameters

startDatestring date

Start date for filtering (YYYY-MM-DD format)

Start date for filtering

endDatestring date

End date for filtering (YYYY-MM-DD format)

End date for filtering

categoriesstring

Comma-separated list of category names to filter by

Categories to filter by (comma-separated)

minAmountnumber

Minimum amount to filter by

Minimum amount to filter by

maxAmountnumber

Maximum amount to filter by

Maximum amount to filter by

accountType'BANK' | 'CREDIT' | 'INVESTMENT' | 'LOAN'

Account type to filter by

Account type to filter by

accountSubtype'CHECKING_ACCOUNT' | 'SAVINGS_ACCOUNT' | 'CREDIT_CARD' | 'PAYMENT_ACCOUNT'

Account subtype to filter by

Account subtype to filter by

includeStatusstring

Comma-separated list of transaction statuses to include (POSTED, PENDING)

Transaction statuses to include (comma-separated)

format'raw' | 'structured'

Response format: 'raw' returns unprocessed transaction data (default), 'structured' returns organized data with groupings, summaries, and category breakdowns.

Response format type (default: raw)

clientMessagestring

Natural language message to apply intelligent filtering using LLM. When provided, filters transactions based on the intent expressed in the message (e.g., 'show me food expenses', 'transactions over 100 reais', 'payments to supermarkets').

Intelligent filter message for LLM-powered transaction filtering

Response

List of transactions (format depends on 'format' parameter)

successboolean
countnumber

Total number of raw transactions

totalBeforeFilternumber nullable

Total number of transactions before intelligent filtering (only present when clientMessage is used)

clientMessageUsedstring nullable

The client message used for intelligent filtering (only present when clientMessage is provided)

messagestring

Descriptive message about the filtering results and processing applied

timestampstring date-time

Example response

{
  "success": true,
  "count": 150,
  "totalBeforeFilter": 250,
  "clientMessageUsed": "show me food expenses over 50 reais",
  "message": "Encontradas 150 transações (de 250 totais após aplicar filtro inteligente baseado na sua mensagem). Os dados foram estruturados para facilitar a análise.",
  "filters": {
    "clientMessage": "show me food expenses over 50 reais"
  }
}