How the bank statement analyzer works

The product is built around one pipeline: ingest the statement, recover the structure, extract the transaction rows, categorize them, then turn that same dataset into a report or an export file.

Upload my statementSee the five steps
statements analyzed
1,000+
transactions categorized
119,000+
balance checked every time
To the cent
to ask for your money back
14 days

From PDF to report in five steps

Every statement goes through the same pipeline, whether it is a digital PDF or a phone photo.

  1. Step 1

    Upload a PDF or image statement

    Users upload a bank statement PDF, JPG, or PNG. The product accepts digital statements, scanned documents, and photographed pages.

  2. Step 2

    Detect text vs scanned layout

    If the statement already contains machine-readable text, we preserve that structure. If it is image-based, OCR runs first so transaction rows can be reconstructed.

  3. Step 3

    Extract rows and normalize fields

    The pipeline separates dates, merchant descriptions, debit or credit amounts, balances, and multi-line entries into clean transaction objects.

  4. Step 4

    Categorize spending with AI

    Each transaction is assigned to a category such as groceries, transport, housing, subscriptions, fees, or transfers, based on merchant and context.

  5. Step 5

    Generate analysis and exports

    The same extracted dataset powers the dashboard, spending breakdowns, recurring charge detection, and downloads such as CSV, Excel, QIF, OFX, QBO, or JSON.

One engine behind every tool

The same reading step powers the analyzer, the scanned-statement reader and the exports, so a statement that works in one works in all of them.

Bank statement analyzerSpending analysisTransaction categoriesScanned statements

Where users should double-check output

  • Very low-quality scans or photos with blur and shadows
  • Statements with handwritten notes or stamps over transaction rows
  • Exports where the bank omits balances or uses unusual debit-credit layouts
  • Rare merchants that require manual category override after the first pass

Questions & answers

Does the workflow differ for scanned statements?

Yes. Digital PDFs skip OCR and usually produce the highest extraction accuracy. Scanned or photographed statements go through OCR first, then the same extraction and categorization pipeline.

What data is extracted from the statement?

The pipeline extracts transaction dates, descriptions, signed amounts, balances where available, debit or credit direction, and an AI-assigned category.

Is this only for analysis, or also for conversion?

Both. The same extraction layer is used for the analyzer dashboard and for file conversion flows like bank statement to CSV, Excel, JSON, OFX, QIF, and QBO.

Something else on your mind? Write to contact@mybankstatementanalysis.com.

Try it on your own statement

Your first upload is free, with no account or card. If a paid report doesn't match your statement, you get your money back within 14 days.

Upload my statement