Accounting API Pricing for Lenders in 2026

Accounting data APIs are priced per borrower connection, not per loan, and that single fact decides whether they are cheap or expensive for you. Merge publishes $650 a month for up to 10 linked accounts and $65 per account after that. Apideck publishes $599 a month for 25 active consumers. Codat and Rutter publish nothing and quote a platform fee plus a per-connection charge after a sales call. Plaid lists a pay-as-you-go tier with no rate. Every figure below was checked against the vendor's own pricing page, Capterra or G2 on 10 September 2026, with the billing unit labeled on each row, because the unit is the part that matters.

Accounting API pricing for lenders, compared

These are the vendors a US business lender actually meets when shopping for live access to a borrower's books. Where a vendor publishes no price, the row says so rather than repeating a rumor.

Two things stand out. Only two of the five accounting API vendors publish a price at all, and the one built most specifically for lending, Codat, is one of the two that does not. If you want the full vendor-by-vendor picture rather than the rates, the Codat alternatives for lenders comparison covers what each one returns and who it is built for.

What does an accounting API cost per funded loan?

Per-connection pricing looks cheap until you divide it by loans, because you pay for every applicant who connects and fund only some of them. Take Merge's published rate as the yardstick, since it is the only complete one: $650 a month covers 10 linked accounts, so the marginal rate is $65 a connection. A lender that gets 100 applicants a month to connect pays roughly $650 plus 90 times $65, about $6,500 a month. Fund 30 of those and the accounting data cost per funded loan is about $217. Fund 15 and it is $433. Codat and Rutter add a platform fee on top of the per-connection rate, so their per-loan figure is higher at low volume and only converges at scale, which is why neither publishes it.

Against that, the MBA's Q3 2025 benchmark put independent mortgage banks at a pre-tax net production profit of $1,201 per loan originated. Small business lenders do not have a comparable public figure, but the arithmetic is the same: a data cost that runs $200 to $400 per funded loan is a real line item, not a rounding error, and the approval rate is the variable that moves it most. That is the same conclusion our bank data API pricing guide reached for the bank-connection side.

Why is the platform fee the number to ask about first?

Because it is the part of the quote that does not scale down. Merge's blog on Codat pricing, which is a rival's account and should be read that way, describes Codat's model as a fixed yearly platform fee for access plus a variable charge per established connection. One integration agency's write-up on the hidden costs of unified accounting APIs estimates Codat platform fees at roughly $12,000 to $24,000 a year with $30 to $50 per connected company per month on top, and Rutter at $25 to $50 per connection per month. Those are the agency's estimates, not vendor rates, and nothing on either vendor's site confirms or denies them; they are useful only as a reminder of the shape of the quote. A community bank connecting 20 borrowers a month is paying the same platform fee as a neobank connecting 2,000, and the per-loan cost at 20 connections can exceed the loan's own origination fee.

Ask any sales-led vendor for four numbers as separate lines: the annual platform fee, the per-connection rate, whether an applicant who starts but never completes the connection is billed, and what a connection costs once the loan is funded and you only want monitoring. Codat's business lending page promotes continuous portfolio monitoring, which is genuinely valuable, but monitoring means the connection stays billable for the life of the loan.

What happens to the applicants who never connect?

This is the cost the per-connection model hides, and it is not a small one. An accounting API only returns data for borrowers who complete the authorization flow into QuickBooks Online, Xero, Sage or NetSuite. Borrowers on QuickBooks Desktop need an on-premise connector installed by someone at the business. Borrowers whose books live with an outside bookkeeper do not hold the login. Landlords, contractors and many sole proprietors have no accounting software at all. Owners who dislike granting a lender live access decline on principle. None of those applicants cost you a connection fee, but every one of them still has to be underwritten, and they arrive as bank statement PDFs, tax returns and an exported P&L.

So the honest total cost of an accounting API is the API bill plus whatever you spend decisioning the unconnected share by hand. If that share is 40 percent of applications, and for many US small business lenders it is, then the API is solving a little over half the problem. The document half is a flat subscription: our bank statement analysis software computes deposits, recurring revenue, NSF and negative days, existing debt payments and loan stacking flags from the PDFs from $99 a month, and the same output is available through the bank statement analysis API for teams that want it in a decisioning flow. Plenty of lenders run both: the API for connecting borrowers, the document tool for everyone else.

Is Codat cheaper than Merge?

Nobody outside a Codat contract can say, because Codat publishes nothing. What can be said is that Merge's published price is the transparent option and Codat's Lending solution is the more lending-specific product. Merge returns standardized accounting data across QuickBooks, Xero, NetSuite and the rest, plus HRIS, CRM and ATS categories a lender does not need. Codat returns categorized bank transactions with B2B merchant categories, a liabilities report that identifies existing loans and credit lines, standardized financial statements and, on premium tiers, a credit model, which is closer to what an underwriter wants. If the lending features matter to you, get the Codat quote and compare it to Merge's $650 plus $65 per account at your real connection volume. If they do not, Merge is the cheaper known quantity.

Do I need an accounting API at all?

You need one if you are a digital lender or card issuer with engineers, your applicants are software-using small businesses who will connect at application, and you want monitoring after funding. In that model a live connection beats a stale PDF every time. You probably do not need one if most of your files arrive as documents, you have nobody to build and maintain the integration, or you underwrite from computed metrics rather than raw categorized data. Codat's own docs say its Lending solution is best for digital lenders, neobanks, corporate card providers and commerce software, which is an honest description of who it fits. A community bank, CDFI, equipment finance company or private DSCR lender is usually better served by an application than an endpoint.

One more integration cost hides here. Even with an accounting API in place, the data still has to land in your LOS or core, and that plumbing is a separate build. Teams without a platform group often route it through a data integration platform rather than writing and maintaining the connectors themselves, which is worth pricing alongside the API rather than after it.

How to compare accounting API quotes before you sign

Run every quote through the same five questions. What is the platform fee, and is it charged even in a month with no connections? What is the per-connection rate, and does it fall at volume? Is an incomplete connection billed? What does a connection cost in monitoring mode after funding? And what share of applicants like yours actually completes the connection, with a reference lender you can call? The last one is the number the sales deck never volunteers, and it is the one that sets your real cost per funded loan.

Figures verified September 2026 against vendor pricing pages, Capterra and G2. Pricing in this category changes often and two of the five vendors publish nothing, so confirm current rates with each vendor before you commit. If your gap is the borrowers who never connect, see how bank-connection alternatives for lenders handle the same problem on the banking side, or upload a real statement above and see what a document-based analysis returns in the time it takes to book a sales call.