What a demo bank feed taught the bookkeeper's keyword rules

Of six open-source accounting projects, only Bigcapital shipped a bank-feed sample; 121 rows read cleanly and showed what the built-in keyword rules miss.

Tools and projects · 1 min read · Source
ai-bookkeeperbookkeepingevaluation

Open-source accounting projects rarely ship real bank exports, so we swept six of them (Bigcapital, LedgerSMB, Frappe Books, ERPNext, Odoo and Firefly III) for statement samples. Only Bigcapital had one: 121 rows of US bank-feed descriptors, such as PURCHASE AUTHORIZED ON 05/27 ROYAL FARMS ....

AI Bookkeeper read the file without a fault. Its built-in keywords suggested an account for 43% of the lines before the run. We added rules for the names that are plainly one thing (Albertsons, Food Lion, Denny's, PayByPhone and a few more) and left Walmart, Venmo and the IRS blank, since the right account depends on what was bought. That is the design: suggest when sure, queue the rest for review, and learn from the corrections.

Try it with your own statement at https://marisstella.app/ai-bookkeeper/ (no account, no key).

More in Tools and projects

All writing · All in this topic · All tools and research