Hi HN — I built Vibooks after paying about $2,000/year for bookkeeping in my company’s first year.

I realized that a lot of SMB bookkeeping is repetitive and rule-based: receipts, invoices, statements, reconciliation, and producing reports for tax filing. There is still accounting judgment involved, but a large part of the work is manual checking and data entry.

That got me thinking about whether AI agents could handle most of this workflow.

The part I didn’t want was to just send documents to an LLM and trust whatever accounting output came back. Bookkeeping needs structure and consistency.

So Vibooks splits the problem in two:

- the app provides the bookkeeping core: double-entry, ledgers, journals, reconciliation, and reports

- AI provides the flexibility: reading files, suggesting classifications, and operating the system through HTTP APIs

Vibooks does not ship with its own agent. I designed it to work with existing general-purpose agents, since tools like OpenClaw, Claude Code, and Codex are already capable enough to operate this kind of workflow.

It’s a local-first bookkeeping app for SMBs, especially for post-period bookkeeping workflows. It’s not tax filing software, and it doesn’t replace an accountant.

Happy to answer questions about bookkeeping with AI agents.