Skip to main content
Version: 1.1


API Endpoints

This API reference provides detailed information about Fraudio's API endpoints and how to interact with them.

To interact with our Rules Management API, as mentioned in the Quickstart, you can send HTTPS requests to our central API URL. You can send these requests to the Fraudio root URL ( with the route /v1/rules:

The Rules Management API is comprised of several endpoints:

What are Rules?

In the context of the Payment Fraud Detection (PFD) product, rules serve as the definitive guiding criteria that can modify the machine learning-derived score of a transaction. These are customer-defined measures that can potentially override the AI's color recommendation based on specific conditions that you set.

Each rule consists of a trigger — a logical expression that matches certain properties of an incoming transaction. For instance, a trigger might be:

transaction.merchant == "suspicious-merchant" && (transaction.mcccode == "1234" || transaction.mcccode == "5678")

Upon the triggering of a rule, a predefined action is taken, such as:

  • Allow: The PFD will suggest a green (good) recommendation.
  • Review: A yellow (cautionary) recommendation is returned, indicating the need for possible investigation or review.
  • Deny: The PFD will return a red (bad) recommendation.

Rules are assigned a priority level, ranging from 1 (highest) to 5 (lowest), to ensure that the most critical rules are evaluated first. In scenarios where multiple rules might be triggered by a single transaction, the rule with the higher priority will determine the outcome. If two rules share the same priority level, the rule that was created first will take precedence, ensuring consistent rule evaluation and action execution.

For a comprehensive list of fields available for defining rule triggers, check out this link.

Every rule defined within the system possesses a status that dictates its current operational state:

  • Enabled: A rule with this status is live and actively evaluates incoming transactions. If its trigger conditions are met, the rule will take the predefined action.

  • Disabled: This status ensures that the rule remains defined within the system but is not actively evaluating transactions. Essentially, while the rule remains in place, it's in a dormant state and will not influence any transaction evaluation.

  • Archived: An archived rule is one that has been "soft deleted." While it will no longer evaluate or influence transactions, it is retained within the system for historical reference and potential future reactivation.

Rules can be seamlessly transitioned between these statuses based on the evolving needs of the fraud detection strategy. The flexibility of these status options ensures that rules can be optimally managed and utilized based on real-time requirements.