Skip to main content
The Reconify engine compares two or more transaction sources, explains what matched and what did not, and produces machine-readable results for reconciliation, accounting, and payment-processing workflows. It’s a local and deterministic engine written in Go and built to process thousands of transactions per second from Excel, CSV and JSON files. Here is why you should use it:
  • Fast. The Go engine indexes and streams instead of loading everything into memory, so it processes thousands of transactions per second and scales to files with tens of millions of rows. See Handle large files.
  • Audit-ready. --audit attaches run provenance to every result: file hashes, timestamps, and the exact config applied. --deterministic makes reruns byte-identical.
  • Built for agents. Reconify exposes versioned, machine-readable commands for inspecting data, discovering capabilities, running reconciliation, and interpreting results. Coding agents such as Claude Code and Codex can operate the engine without parsing terminal prose or guessing which matching behavior is available. See Run it with an agent.
Reconify does not stop at a matched/unmatched decision. It classifies the relationship between records – including exact matches, amount differences, timing differences, unmatched records, and duplicates – and preserves the evidence needed to explain why each outcome was produced.

Quickstart

Run it with an agent

Matching algorithm