Fraud Protection
Fake and duplicate orders waste delivery fees and agent time. Kllix scores every storefront order in real time, before it reaches your team or your carrier.
How scoring works
When an order is submitted, Kllix reads its context — IP address, phone, device ID, time spent on the page and cart. Rules are evaluated in three layers, and each triggered rule adds to a risk score.
| Score | Decision | Result |
|---|---|---|
| 0–39 | Allow | Order is created normally |
| 40–69 | Flag | Order is created and saved to suspicious orders |
| 70–99 | Challenge | Customer must pass a CAPTCHA |
| 100+ | Block | Order is rejected and logged |
Thresholds are configurable per store.
Layer 1 — Hard blocks
- IP blocklist — block known bad IP addresses.
- Country restriction — only accept orders from the countries you sell to.
- Honeypot — an invisible field that only bots fill in.
Layer 2 — Rate limits
Limit how many orders one source can place per hour and per day:
- Per IP address
- Per phone number
- Per device
Layer 3 — Risk signals
- Fast submission — the form was filled faster than a human could.
- Repeat order — the same customer ordered the same product recently.
- Device ↔ phones — one device used with many phone numbers.
- VPN / proxy — the connection comes from a known VPN or proxy.
- Bot behaviour — automated interaction patterns.
CAPTCHA
Challenged customers see a CAPTCHA. Kllix supports reCAPTCHA v3 and Cloudflare Turnstile — configure your keys in **Settings → Security**.
Configure rules
Go to **Settings → Fraud protection** to enable or disable each rule and adjust its score and limits. Review flagged and blocked attempts, with the rules that triggered them, in **Orders → Suspicious**.