Bank reconciliation

Journal-bank reconciliation

Reconcile bank movements to journal or ledger rows.

Codex handles changing customer formats and mapping judgment. Deterministic Python scripts normalize rows, match evidence and preserve the audit trail.

What it does

From bank and journal files to reconciliation evidence

Use it when the journal-bank reconciliation is structurally different from open-item reconciliation: bank statement rows must be matched directly to journal or ledger rows.

input/
  bank/
    statement.xlsx
  journal.xlsx
  sample_movements.xlsx
out/
  inspection.json
  suggested_recipe.json
  reconciliation/
    normalized_bank.csv
    normalized_journal.csv
    reconciliation_matches.csv
    unmatched_bank.csv
    unmatched_journal.csv
    journal_bank_reconciliation.xlsx
    reconciliation_audit.json
    review_notes.md

Inputs

What it reads

The plugin reads local Excel, CSV and text-PDF inputs. OCR-only PDFs are flagged for review rather than treated as reconciled.

Outputs

What it produces

The result is a reviewable work folder: mapped recipe, normalized bank and journal rows, match table, exception tables, workbook and audit JSON.

Reviewable reconciliation, guided in Codex

Codex helps inspect the customer format, propose mappings and ask only for unresolved choices before the reconciliation files are produced.

If a customer format fails, the next improvement is a tested parser or mapping rule, not duplicated web-app logic.

Prompts

Starter prompts

Start from one of these prompts and replace the file and folder paths. Codex will ask before resolving essential ambiguities.

Download

Download the plugin

Requires Codex desktop. One ZIP installs the standard Accounting Plugin Pack.

Installation

Official download: Codex desktop.

    First use

    In a Codex chat, provide the bank path, journal path, language and thresholds.

    Use Journal-Bank Reconciliation on bank statements in /path/bank and journal file /path/journal.xlsx.
    Language: en. Document language: auto.
    Inspect columns and files, ask only for essential ambiguities, then generate reconciliation, diagnostics and audit trail.