Offers

Codes that leak
and get reused

Catch code guessing, repeat redemptions and offers spreading across coupon sites.

Decision trace

Example · default thresholdsLeaked code
  1. InputThe same code redeemed from ten IP prefixes in one hour.
  2. Reasonpromo_leaked
  3. Decisionlimit · one promo.abuse webhook queued for the code

Before and after redemption

  1. Gate the offer

    Your backend sends the code with result attempt. Check the decision before applying a discount.

  2. Report the outcome

    Record redeemed, invalid or rejected. Five invalid codes per device or twenty per IP prefix block further attempts for an hour.

  3. Review linked use

    Two campaign codes on one device or linked-account cluster trigger a limit. One code used across linked accounts also triggers a limit.

The dashboard shows leaked code hash prefixes, campaigns and counts, plus top enumerating sources. Adjust thresholds or clear history from site settings. Raw codes are never stored.

Templated email addresses

Templated email addresses reveal another kind of repeat signup. Five similar locals on one domain in 24 hours can add email_pattern_burst, using a narrow numeric range or shared four-letter prefix. Freemail needs the prefix match: john.smith1984 and jane.doe1990 on Gmail are not grouped by their years alone.

Random-looking locals are telemetry only. They carry no score or repeat-limit weight.

API

const result = await nolgard.promo({
  code: request.body.code,
  accountId: user.id,
  ip: clientIp,
  campaign: 'fall',
  result: 'attempt'
});
// Apply your offer only when the decision permits it.
// Afterward report redeemed, invalid, or rejected.

Run this server-side. A browser token minted for action redeem adds device identity. Leak detection sees how redemptions spread; it never reads coupon websites.

Limits

Related