PCI Vault Logo

PCI Vault has 3 tokenization algorithms available. You can specify the algorithm to use when you create a capture endpoint or when you send a PAN to the vault directly.

All of these tokens are based on a number, which we obtain from your data. The number that is used is determined in the following way:

  1. First the submitted JSON object is checked for these fields in order: number, card_number, n, account_number. If a field with one of these names are found, its value is used to generate the token.
  2. If none of these fields are present in the JSON object, the query parameters of the HTTP request are checked for the same fields in the same order.
  3. If none of these fields are available in either the JSON object or in the query parameters, the request fails with status code 400.

Please note that all algorithms are sensitive to spaces. This means that a PAN with spaces will generate a different token than a PAN without spaces.

Full Tokenization (Default)

The default PCI Vault tokens are the most collision resistant of the 3 algorithms. Full tokenization uses a proprietary cryptographically secure hashing algorithm to generate a hash, which then serves as a token.

These tokens are 64 characters long, highly collision resistant, and the recommended algorithm unless you have specific needs.

To use this tokenization algorithm, specify algorithm=full in your request, or send the request without specifying an algorithm. The Full algorithm is the default.

Example tokens:

Number Token
4111111111111111 9bfe9241f7ad0d84230e9837aa99c64bf3c3936478317146d1f7c58a66f40af6
5425233430109903 bf2fa482286633fab50369c8ba855bf7bc5b1d29bb5ffc183b371e20b9d6253d
4263982640269299 744d174e68756188bb48dd69ff69d1a4f4ed3a2ca1d3d0747b3b8d768f010e0b
374245455400126 a8366e50f6aafd1cc71360f43060808f2ef9f6f8a850319aff629a0b80ddd7de

Partial Tokenization

Partial tokenization are useful when you want to restrict the token length to be the same length as the input number. However, the trade-off to achieve this is a significantly lower collision resistance. It is therefore recommended to use a unique reference for every token in order to guarantee uniqueness. Partial tokenization works by generating a Full token, and trimming the resulting token to the desired length. The specified digits of the number is then replaced by the generated token. The resulting token is part original number, and part token.

These tokens are the same length as the input number, but not highly collision resistant.

To use this tokenization algorithm, specify algorithm=partial_n_pos in your request. Where n is a number between 8 and 16 (inclusive) and pos is either start or end

Example tokens for partial_8_end:

Number Token
4111111111111111 411111119bfe9241
5425233430109903 54252334bf2fa482
4263982640269299 42639826744d174e
374245455400126 3742454a8366e50

Luhn Compliant Tokenization

Luhn Compliant tokenization generates a token that is in itself a 16 digit credit card number that also passes Luhn's algorithm.

These tokens are always 16 characters long and always pass Luhn's algorithm. They are less collision resistant than Full tokens, but they are collision resistant enough for most practical purposes. It is recommended to limit the amount of tokens to a few thousand per unique reference to be safe.

To use this tokenization algorithm, specify algorithm=luhn in your request.

Example tokens:

Number Token
4111111111111111 5560386204722122
5425233430109903 3113674056387905
4263982640269299 3543337643660602
374245455400126 3935235944928114