AI-Process Credit Card in n8n

AI-Process Credit Card is a PDF4me node action in n8n that reads a credit card image or PDF and returns the card details as structured JSON fields. Use it to capture cardholder name, card number, expiry, and issuing bank from a scan or photo instead of retyping them into a form by hand.

What this node does

PDF4me: AI-Process Credit Card takes a card file from Binary Data, a Base64 String, or a URL and applies vision AI to return a creditCardData object holding the individual card fields. It accepts PDF, PNG, JPG, and JPEG. Because it returns discrete values rather than a blob of text, the next workflow step can act on a single field, and you can choose deliberately which of those values are worth keeping.

Authenticating Your API Request

Every PDF4me node in n8n requires a valid Credential to connect with. Create or select one that holds your PDF4me API key so the workflow can authenticate extraction requests securely.

Get your API Key

Important Facts You Should Not Miss

Extracting is not the same as storing

Card numbers and security codes are regulated data. Decide field by field what the workflow persists, and drop the rest before anything reaches storage or logs.

Empty strings, not missing keys

Fields the model cannot read come back as "" rather than being absent, so test for empty values instead of checking whether a key exists.

Image quality drives accuracy

Glare, angle, and embossed digits all reduce OCR reliability. Use the warnings array and fallbackUsed to route doubtful results to review.

PDF4me AI-Process Credit Card node in n8n, reading a credit card file and returning structured card fields — PDF4me AI-Process Credit Card action in n8n

Card OCR vs a Generic Document Parser

PDF4me ships both a card-specific action and a general parser. The card action knows what a card looks like; the general parser is told what to look for.

Aspect	AI-Process Credit Card	AI Document Parser
Field set	A fixed card schema returned as creditCardData	Whatever fields you define for the document
Setup effort	None, the card fields are already known	You describe the fields you want extracted
Best fit	Payment cards specifically	Any document shape the card action does not cover
Extending it	Custom Field Keys for extras beyond the card schema	Fully defined by you from the start

What Parameters Does AI-Process Credit Card Need?

Required: Action, Input Data Type, the field matching that input type, and Credit Card Name. Custom Field Keys is optional.

Parameter	Required	What it does	Example
Action	Required	Selects the PDF4me node action to run. Choose AI-Process Credit Card.	`AI-Process Credit Card`
Input Data Type	Required	Format of the source card input. Choose Binary Data (from a previous node), Base64 String, or URL.	`Binary Data`
Input Binary Field	Conditional	Name of the binary property on the incoming n8n item that holds the card file. Required when Input Data Type is Binary Data. Defaults to data.	`data`
Base64 Credit Card Content	Conditional	Base64-encoded content of the card file. Required when Input Data Type is Base64 String.	`JVBERi0xLjQKJ...`
Credit Card URL	Conditional	Publicly reachable HTTPS URL to the card file. Required when Input Data Type is URL.	`https://example.com/card.jpg`
Credit Card Name	Required	Filename of the input file, used for format detection. Accepts .pdf, .png, .jpg, and .jpeg extensions, so give it the extension the file actually is.	`card.jpg`
Custom Field Keys	Optional	List of extra field keys to pull beyond the standard card schema. Leave empty unless the source carries values the built-in fields do not cover.	`["loyaltyId", "branchCode"]`

Output Fields

The node returns JSON rather than a file. The card values live inside the creditCardData object.

Field	Type	What it contains
`creditCardData`	Object	The extracted card fields. See the field breakdown below.
`jobId`	String	Identifier for the AI processing job, useful when reporting an extraction problem.
`jobIdExt`	String	External job identifier for third-party integration, null when unused.
`success`	Boolean	True when extraction completed. Use it to branch error handling.
`message`	String	Human-readable status message, carrying success confirmation or error detail.

Fields Inside creditCardData

Any field the model cannot read on a given image comes back as an empty string.

Field	Type	What it contains
`cardholderName`	String	Name as it appears on the card.
`cardNumber`	String	The primary account number. Regulated data, so handle deliberately.
`expiryDate`	String	Expiry date, typically MM/YY.
`validFrom`	String	Valid-from date where the card shows one.
`validThru`	String	Valid-through date, often the same as expiryDate.
`cvv`	String	Card security code. The PCI DSS standard prohibits retaining this after authorisation.
`cardType`	String	Card type where detected, such as credit or debit.
`cardBrand`	String	Card brand where detected, such as Visa or MasterCard.
`bankName`	String	Bank name printed on the card.
`issuingBank`	String	Issuing bank where it can be determined separately from bankName.
`accountNumber`	String	Associated account number when the source is a statement rather than a card face.
`warnings`	Array	Messages about data quality or extraction difficulty on this file.
`fallbackUsed`	Boolean	Whether an alternative extraction path was used. A useful low-confidence signal.

How Do I Set Up AI-Process Credit Card in n8n?

Add PDF4me to your n8n workflow and choose the AI-Process Credit Card action.
In Credential to connect with, select your PDF4me credential or click Create New Credential and paste your API key.
Set Input Data Type to Binary Data (default), Base64 String, or URL and supply the matching card file field.
Set Credit Card Name with the correct extension, for example card.jpg or statement.pdf.
Optionally add Custom Field Keys for values beyond the standard card schema.
Execute the node and read the creditCardData object from the response.
Before storing anything, select just the fields you actually need and drop the rest.

Typical Setups

Workflow ExamplesCommon n8n workflow patterns using AI-Process Credit Card.

Onboarding without retyping

A customer uploads a photo of their card during sign-up.
AI-Process Credit Card reads the cardholder name and expiry.
Only the non-sensitive fields are used to prefill the form for confirmation.

Route low-confidence reads to a human

The node processes an uploaded card image.
An IF node checks the warnings array and fallbackUsed flag.
Anything doubtful goes to a review queue instead of straight through.

Expiry monitoring on file

A card image is processed once at capture time.
Only expiryDate and the cardholder name are retained.
A scheduled workflow flags cards approaching expiry for renewal outreach.

Match a statement to an account

A card statement PDF arrives in a shared mailbox.
The node extracts bankName and accountNumber.
The statement is filed against the matching account record.

Practical Tips

Drop what you do not need immediately

Select the handful of fields the workflow actually uses in the very next node, so the full card payload never reaches storage, logs, or an error notification.

Give the filename its real extension

Credit Card Name drives format detection. Calling a JPEG card.pdf invites a failed or degraded read.

Check warnings before trusting a value

The warnings array and fallbackUsed are the cheapest confidence signal available. Branch on them rather than assuming a clean read.

Test for empty strings, not missing keys

Unread fields return as empty strings, so a key-existence check will pass on a field that holds nothing.

Improve the input before blaming the model

A flat, evenly lit, high resolution capture beats any amount of downstream correction, especially on embossed digits.

Use Custom Field Keys sparingly

The standard schema covers the card face. Reach for custom keys only when the source genuinely carries something extra.

Cheat Sheet

Field	Value
Action	`AI-Process Credit Card`
Input Data Type	`Binary Data`
Input Binary Field	`data`
Credit Card Name	`card.jpg`
Custom Field Keys	`(empty)`
Accepted formats	`PDF, PNG, JPG, JPEG`
Returns	`JSON with a creditCardData object`
Credentials	`PDF4me API credential`

Common Questions

What file formats can the node read a credit card from?+

PDF, PNG, JPG, and JPEG. Set Credit Card Name with the matching extension so the format is detected correctly, for example card.jpg for a photo or statement.pdf for a document.

Which fields come back, and why are some empty?+

The creditCardData object carries fields such as cardholderName, cardNumber, expiryDate, validThru, bankName, and cvv. Fields the model cannot read on a given image come back as empty strings rather than being omitted, so check for empty values instead of assuming every key is populated.

Should I store the extracted card number and CVV?+

Treat that decision carefully. A primary account number and security code are exactly the data the PCI Security Standards Council governs through PCI DSS, and the standard prohibits retaining the card verification code after authorisation. Extracting a field in a workflow does not mean you should persist it, so drop what you do not need before writing anything to storage or logs, and confirm your own obligations against the current standard.

What are the warnings and fallbackUsed fields for?+

warnings is an array of messages about data quality or extraction difficulty on that file, and fallbackUsed indicates whether an alternative extraction path was taken. Both are useful signals for routing low-confidence results to a human review branch in your n8n flow logic instead of straight through.

How accurate is OCR on a photographed card?+

Accuracy depends heavily on the source image. Embossed numbers at an angle, glare, and low resolution all reduce it. Check the success flag and the warnings array, and route anything doubtful to manual review rather than trusting a single pass on a photograph.

AI-Process Bank Statement

Read transactions and balances when the source is a statement rather than a card.

AI-Process Receipt

The natural companion in an expense workflow, capturing what was actually bought.

AI Document Parser

Define your own field set when the fixed card schema is not the right shape.

AI Invoice Parser

Extract line items and totals from invoices in the same accounts pipeline.