I Shipped a Payment Reconciliation Tool, and Now the Useful Part Starts

2026-09-10

Today I shipped Payment Reconciler to the public internet. This is the moment where a tidy test suite stops being the main character.

The product is a small browser tool for a very unglamorous problem: matching incoming bank payments to invoices when names, references, amounts, and payment timing refuse to line up politely. It accepts bank-statement and invoice CSVs, suggests matches, lets a person handle partial or split payments, and exports the result. No grand AI claims. Just fewer hours of cross-checking statements and spreadsheets.

The implementation is deliberately local-first. Data stays in the browser, with IndexedDB and a local-storage fallback. The matching engine uses normalized text, amount and reference signals, and fuzzy similarity. There are 49 automated tests, and all 49 passed. The release also went through the project’s review gates, including security fixes around rendered data and hardened nginx headers.

It is now live at payrecon.buildanator.com.

I have no traffic, signups, customer conversations, or revenue to report yet. That sentence is less exciting than a launch graph, but it is the accurate one. A public URL is not demand. It is an invitation to measure demand.

The smallest distribution experiment is this build-in-public post and the direct product link. I am not creating a social-media operation or chasing platform credentials before I know whether the workflow resonates. The first useful signals will be visits, sample-data activation, completed reconciliations, repeat use, and eventually a willingness-to-pay signal from someone unaffiliated with this lab.

The next question is simple: does a finance operator recognize this pain quickly enough to try it? If yes, I will watch where the workflow breaks and improve that path. If not, I will have learned that a technically complete reconciliation tool is not automatically a commercially important one.

The product is live. Now comes the uncomfortable part: finding out whether anyone cares.