Skip to main content
Reading a client’s transactions takes two calls. GET /api/v2/account/{username} lists the client’s banks, with each bank’s source and the id of each account. GET /api/v2/transactions/list then returns the transactions LedgerSync has stored for an account, newest first, one page at a time.
Base URL: https://ledgersyncappv2.com  ·  Sandbox: https://staging.ledgersyncappv2.com, same paths, with its own sign-up at https://staging.ledgersyncappv2.com/sign-up  ·  Auth: sign in with POST /api/v2/public/authenticate and send the authToken it returns as the Auth-Token header on every call. See v2 authentication.

How to call it

1

List the client's banks and accounts

Put the client’s username in the path. It is the username GET /api/v2/client/list returns.
Response (trimmed)
Each entry in banks is one bank connection. Keep two values from it: the bank’s source (FINICITY, MX, FDE or PDF) and the id of each account in accounts.
2

List each account's transactions

For every account, send the bank’s source in uppercase and the account’s id as accountId:
source defaults to FINICITY. Leave it out for an MX account and the call reads FINICITY data instead: it answers 200, without that account’s transactions and without an error. Always send the source of the account’s bank.
A PDF bank cannot be read here: source takes FINICITY, MX or FDE only.
3

Read every page

page starts at 0. Ask for the next page until one comes back with fewer rows than rowsPerPage (50 unless you set it). See Paging.
4

Store each row under its transactionId

Upsert on transactionId, so reading the same days again updates the rows you already have instead of adding copies. See Choosing a transaction key.

Example: a FINICITY account

  • Dates are yyyy-MM-ddTHH:mm:ss in UTC with no offset. Read the first 10 characters as the date.
  • bankAccountId in a FINICITY row is the provider’s account id, the same value as accounts[].accountId in the account list.
  • FINICITY pending rows have postedDate: null, so this request, which sends from and to, leaves them out. See Dates.

Example: an MX account

  • Dates are MM/dd/yyyy.
  • The PENDING row has no posted date yet, so its postedDate holds its transaction date.
  • transactionHash is always null on MX rows.
  • bankAccountId in an MX row is the LedgerSync connection id, the same value as banks[].id in the account list.

Parameters

Without accountId or bankAccountId, the response holds every account of the client for that source. An accountId that is not one of this client’s accounts for that source returns 200 with no rows. Parameters this endpoint does not define, such as id, are ignored. What each parameter does depends on the source:

Dates

  • FINICITY and FDE send postedDate and transactionDate as yyyy-MM-ddTHH:mm:ss in UTC with no offset, for example 2026-09-28T12:00:00. Read the first 10 characters as the date.
  • MX sends MM/dd/yyyy, for example 09/28/2026.
  • postedDate is the day the bank posted the transaction. transactionDate is the day it happened, for a card the purchase date. Both are the bank’s own values: LedgerSync does not adjust them.
  • from and to are both included.
  • FINICITY pending rows have postedDate: null. A request with from or to leaves them out. Omit both dates to include them.
  • MX pending rows carry their transaction date in postedDate, and for MX from and to filter on the transaction date, so they stay in range.

Paging

  • page starts at 0. rowsPerPage is 10, 20 or 50, default 50.
  • totalItems counts every matching transaction across all pages. selectedPage is the page you asked for.
  • A page past the end returns an empty transactions list with the same totalItems.
  • To read everything, request page 0, 1, 2 and so on until a page holds fewer rows than rowsPerPage.

Broken connections

The transaction list reads only what LedgerSync has stored and never contacts the bank. A broken connection still returns every transaction collected before it broke, and one connection’s status never affects another connection’s transactions. Connection health is in the account list: A FAILED connection keeps its stored transactions, and the transaction list keeps returning them.

Ids

Because ids are numbered separately per source, the same number can name a FINICITY account and an MX account. Sending the wrong source returns no rows, or the rows of a different account that has that id under that source. Always send the source of the account’s bank.
The bankAccountId request parameter is always a connection id (banks[].id). The bankAccountId field in a transaction row is not: it means something different per source, as the table shows.

Transaction fields

FINICITY type values: atm, cash, check, credit, debit, deposit, direct debit, direct deposit, dividend, fee, interest, other, payment, point of sale, repeat payment, service charge, transfer. Rows also carry accountNumber, normalizePayeeName, categorization, payee, memo, currencySymbol, checkNum and runningBalance. Any of them can be null.

Choosing a transaction key

Store transactionId. It joins the source and the id, so it is unique across sources. A refresh updates the stored row in place, so a FINICITY or MX transaction keeps its transactionId for as long as the bank keeps reporting it.
  • If you store id instead, store source with it: FINICITY 119874 and MX 119874 can both exist.
  • Do not use transactionHash as the key. It is built from the account’s last 4 digits, the transaction date, the first 30 characters of the description, the amount and the type, so two identical purchases on the same day share it. MX rows have none, and a FINICITY row has none when its account number has fewer than 4 digits or it has no transaction date.
A transaction can still come back under a new transactionId:

Statuses

FINICITY rows carry active, pending or shadow. MX rows carry POSTED or PENDING. Book settled rows (active, POSTED), hold pending ones, and leave out shadow rows. Transaction status explains each value and how pending rows settle.

Errors and empty answers