Victor Ezenwa@vickman
1
Events attended
1
Submissions made
Nigeria
1 year of experience
About me
I build cool apps
I built with
VULTR🤓 Latest Submissions
Ma'at: Invoice & Payment Reconciliation
Ma'at helps businesses and finance teams reconcile invoices against received payments. Users sign in to a private workspace, upload invoice and payment CSVs, check column mappings, and validate their records before matching begins. Each account starts empty and displays only its own uploaded data. Deterministic rules match clear cases using invoice references, currencies, and amounts. Partial payments, ambiguous candidates, unpaid invoices, and unallocated payments remain visible for human review. Reviewers can inspect original source rows, confirm allocations, choose another eligible payment, reject suggestions, or leave items unresolved. Undo records a reversal without erasing the original decision. Vultr Serverless Inference explains exceptions using a small bundle of relevant evidence. AI responses include source references and limitations, but cannot change allocations or financial totals. Human decisions remain separate from AI advice. The application runs on a Vultr virtual machine, with a Next.js web interface, a separate background matching worker, SQLite storage, and private CSV files. Privy provides email authentication verified by the server. The interface supports desktop and mobile, with downloadable reconciliation reports, unallocated-payment exports, and decision history. Ma'at makes reconciliation easier to review and explain while keeping unresolved items and the evidence behind each decision visible.
8 Nov 2026
