Bank statement parser

Extract bank statement transactions into structured data

Define the fields you need, extract transactions from PDF statements, and export the results to Excel, CSV, or your existing workflow.

Parse your first statement See a sample result

Free trial: 20 credits per month, no credit card required.

Synthetic example: a bank statement and the fields extracted from it

Try it on a sample

A sample statement and the result to expect

Download a two-page PDF statement, upload it to your own inbox, and compare your result with the one below. The statement is complete: every transaction of the period is in the table, and the balances add up (opening balance + deposits − withdrawals = closing balance).

Download sample statement (PDF)Download result (Excel)

Synthetic statement: the bank, the account holder, and all numbers are fictional.
2 pages and 14 transactions: the full period, not an excerpt.
Separate withdrawal and deposit columns, plus a running balance.
The table header repeats on page 2, and some descriptions wrap onto a second line.

Account fields

account_holder_name: Alex Sample
account_number_masked: ****4821
account_type: Checking
bank_name: Lakeside Sample Bank
closing_balance: $8,932.34
currency: USD
opening_balance: $4,250.00
statement_period_end: 08/31/2026
statement_period_start: 08/01/2026
total_deposits: $6,210.87
total_withdrawals: $1,528.53

Transactions (14 of 14)

transaction_date	transaction_description	withdrawal_amount	deposit_amount	transaction_balance
08/03/2026	CARD PURCHASE - GREENLEAF GROCERY #114	86.42		4,163.58
08/04/2026	ACH DEPOSIT - BRIGHTWAVE PAYROLL		2,480.00	6,643.58
08/05/2026	CHECK #1042	750.00		5,893.58
08/07/2026	ONLINE TRANSFER TO SAVINGS ****7788	300.00		5,593.58
08/10/2026	CARD PURCHASE - CITY TRANSIT AUTHORITY	42.00		5,551.58
08/12/2026	AUTOPAY - PINE VALLEY POWER	118.37		5,433.21
08/14/2026	WIRE TRANSFER IN - ORCHARD & SONS LLC INVOICE 2291 CONSULTING		1,250.00	6,683.21
08/17/2026	CARD PURCHASE - RIVERSIDE PHARMACY	23.15		6,660.06
08/19/2026	ATM WITHDRAWAL - 210 ELM ST	100.00		6,560.06
08/21/2026	MONTHLY MAINTENANCE FEE	12.00		6,548.06
08/24/2026	ACH DEPOSIT - BRIGHTWAVE PAYROLL		2,480.00	9,028.06
08/26/2026	AUTOPAY - NORTHLINK INTERNET	64.99		8,963.07
08/28/2026	CARD PURCHASE - HARBOR BOOKS	31.60		8,931.47
08/31/2026	INTEREST PAID		0.87	8,932.34

Result: what Airparser returned for this sample statement, shown unedited; the download is the Excel export of the same run. Field names and formats come from the schema used for this run; yours follow the schema you define. This is one example run, not an accuracy benchmark.

Set up an inbox with fields like these and upload the sample PDF, or go straight to your own statement.

Parse your first statement

From PDF to CSV

How to parse a bank statement with Airparser

Upload or email a bank statement PDF. Define fields for account details and transactions. Review and edit extracted values, then export to Excel, CSV or JSON, or connect Google Sheets. Reuse your inbox and schema for future statements.

Illustrative workflow using sample data. Review extracted values before export.

Parse your first statement

1. Create an inbox
Name it and click Create. The inbox is where statements arrive and where your field definitions live. You can leave the parser engine on its default, Vision.
2. Upload and parse
Drop a PDF into the inbox, or forward it by email to the inbox address. Airparser can set up the fields and parse it automatically, or you can define your own fields for more control.
3. Export or send it anywhere
Export XLSX, CSV, or JSON, or send the data anywhere: Google Sheets, webhooks, the API, Zapier, Make, or n8n.

In the app

See it in the app

Statements in PDF are hard to reuse: copying tables into Excel can break the column alignment. In Airparser the original document stays next to the parsed fields, so you can check each value against the source.

These screenshots show a public sample statement processed in Airparser, with the transactions extracted as a table and available as JSON.

Review values side by side with the original
Export the transactions to Excel or CSV
Send rows to Google Sheets
Push results to your own systems with webhooks or the API

The statement next to the parsed fields

Airparser showing a sample bank statement PDF next to its parsed fields and transactions table

The same result as JSON

JSON view of the extracted bank statement fields and transactions in Airparser

Vision engine for complex layouts

The default parser engine in a new inbox is Vision, which we recommend for statements with complex layouts or multi-column tables. You can switch to the Text engine in the inbox's advanced settings.

Every statement is different, so check the results on your own documents before relying on them. See the quality notes below.

Exports are available as XLSX, CSV, or JSON. For a recurring flow, connect Google Sheets, a webhook, or an automation platform such as Zapier, Make, or n8n.

Learn more about converting PDF bank statements to Excel or CSV →

Regular use

Set it up once, then keep the statements coming

Fields belong to the inbox, so every new statement is parsed against the same definition. What happens next depends on where the data needs to go.

Built into Airparser

Upload PDFs in the app, or forward them by email to the inbox address.
Export XLSX, CSV, or JSON on demand.
Google Sheets integration: each transaction row becomes a spreadsheet row.
Webhooks: send the parsed data to your server or another platform.
REST API for your own code.

Through Zapier, Make, n8n, or your own code

Route parsed data to other apps, such as accounting tools, CRMs, or databases, using the connectors of these platforms.
The destination side is configured in the platform, not in Airparser, so test the workflow with the sample statement before running real ones.

Quality

Check the results on your own statements

Statements vary by bank, layout, and scan quality, so we don't publish a single accuracy figure. These are the places worth checking first:

Scan quality

Blurry, skewed, or low-resolution scans are harder to read. When you can, start from the PDF your bank generated rather than a photo or scan.

Wrapped descriptions

A long description that continues on a second line belongs to one transaction. Check that no wrapped line turned into a separate row.

Amounts and signs

Confirm that withdrawals and deposits land in the right column and that decimal and thousands separators were read correctly.

Page breaks

Check the rows around page breaks, and make sure a repeated table header did not create an extra row.

Balances reconcile

Opening balance + deposits − withdrawals should equal the closing balance. If it does not, a row is missing or misread.

If a layout gives you trouble, tell us at [email protected].

Credits and data handling

How credits work

Airparser uses credits: 1 credit per PDF page, so a two-page statement uses 2 credits. The free trial includes 20 credits per month, no credit card required.

See pricing

Your statements and your data

Documents you send are not used to train AI models. You can delete any document at any time, and retention can be set from 1 to 180 days (the free trial keeps documents for 30 days).

Trust Center AI & data security

What you can extract

Fields you define

Account information

Account number, bank name, account holder, statement period

Transaction details

Date, description, amount, reference or check number

Withdrawals & deposits

Separate withdrawal and deposit amounts, or a single signed amount, depending on how your statement is laid out

Balances

Opening balance, closing balance, running balance

Your own fields

Add any field your workflow needs and choose its type, such as text, number, date, or table

Result for the sample statement (first 3 of 14 transactions):

JSON

{
  "account_holder_name": "Alex Sample",
  "account_number_masked": "****4821",
  "account_type": "Checking",
  "bank_name": "Lakeside Sample Bank",
  "closing_balance": "$8,932.34",
  "currency": "USD",
  "opening_balance": "$4,250.00",
  "statement_period_end": "08/31/2026",
  "statement_period_start": "08/01/2026",
  "total_deposits": "$6,210.87",
  "total_withdrawals": "$1,528.53",
  "transactions": [
    {
      "deposit_amount": "",
      "transaction_balance": "4,163.58",
      "transaction_date": "08/03/2026",
      "transaction_description": "CARD PURCHASE - GREENLEAF GROCERY #114",
      "withdrawal_amount": "86.42"
    },
    {
      "deposit_amount": "2,480.00",
      "transaction_balance": "6,643.58",
      "transaction_date": "08/04/2026",
      "transaction_description": "ACH DEPOSIT - BRIGHTWAVE PAYROLL",
      "withdrawal_amount": ""
    },
    {
      "deposit_amount": "",
      "transaction_balance": "5,893.58",
      "transaction_date": "08/05/2026",
      "transaction_description": "CHECK #1042",
      "withdrawal_amount": "750.00"
    }
  ]
}

Overview

What is a bank statement parser?

A bank statement parser reads statements, usually PDFs, and turns the account details and transaction lines into structured data you can sort, reconcile, and import. Instead of retyping or copy-pasting tables into Excel, you get rows and columns.

Airparser uses LLM and vision models to read the document, and you decide which fields to extract. Typical fields for a bank statement:

Account numbers and bank details
Statement period
Transaction dates and descriptions
Withdrawal and deposit amounts
Opening, closing, and running balances
Check numbers and reference codes
Customer name and address
List of transactions with full details

What our customers say

What Airparser users say

The Right Product for quick integration of smart extraction of data from PDF.
Marty N.
CEO, Information Technology
on Capterra

The simplicity of creating the data capture fields. I also like the webhook feature.
Allen M.
Owner Realtor, Real Estate
on Capterra

The AI schema creator made it quite literally a 60 second job.
NK
Nick K.
Family Owner, Retail
on Capterra

Bank Statement Parsing Use Cases

Accounting & Bookkeeping

Turn statement lines into rows you can reconcile against your books.

Financial Analysis

Extract transaction data for cash flow analysis, expense tracking, and financial reporting.

Loan Applications

Extract balances and transactions from applicants' statements for credit review.

Compliance & Audit

Extract transaction data for compliance reporting and audit purposes.

System Integration

Send parsed rows onward: Google Sheets, webhooks, the API, or Zapier, Make, and n8n.

KYC & Verification

Extract account information and transaction history for KYC processes and identity verification.

Related parsers

Credit Card Statements

Cardholder, payment, and transaction extraction for expense management.

PDF Table Parser

Extract the table columns you need from other PDF documents.

Frequently Asked Questions

How accurate is the bank statement parsing?

Accuracy depends on the statement's layout and scan quality, so we don't quote a single number. Try it on your own statements (the free trial includes 20 credits per month), then check the rows: opening balance + deposits − withdrawals should equal the closing balance, and you can correct any value with Edit data.

What file types can I upload?

PDFs, including scanned PDFs, and images (JPG, PNG, BMP). Results on scans and photos depend on image quality, so a PDF generated by your bank usually gives the cleanest input.

Can I convert PDF bank statements to Excel or CSV?

Yes. Create an inbox, define the fields you want (account details and a table for the transactions), upload the statement, and export the parsed data as XLSX, CSV, or JSON. You can also send transaction rows to Google Sheets, or onward through webhooks, the API, or Zapier, Make, and n8n.

How are credits used for bank statements?

One credit is used per PDF page, so a two-page statement uses 2 credits. The free trial includes 20 credits per month, no credit card required.