Fraud and chargebacks

Fraud Shield. Ban Radar. Chargeback Shield.

Fraud Shield refuses the buyer before the payment form loads. Ban Radar watches your dispute rate against the 0.75 percent line Stripe uses, and Chargeback Shield files the evidence before the deadline runs out.

One order, three checkpoints

Fraud Shield

Order created

  • 16 block types
  • Risk score 0 to 100
  • Per product ceiling

Refuses the order

Ban Radar

Payment session

  • 3D Secure forced
  • VPN and proxy blocked
  • 5 orders per IP a day

Tightens checkout

Chargeback Shield

Dispute opened

  • Evidence packet built
  • Visa CE 3.0 attached
  • Sent before the deadline

Answers the bank

Fraud Shield

Fraud Shield scores every order 0 to 100.

The score comes from the network the buyer connected over, and you set the ceiling per product, so your cheap presets and your lifetime licenses never have to share one rule. Under the score sits a blocklist that reaches the network and the card.

Down to the network
ASN, ISP, city and country, so a blocked buyer cannot just reconnect and try again.
Device and card
Both fingerprints survive a new email address and a fresh signup.
Four switches take no value
Turn on VPN, Tor, hosting or disposable email and every one of them is refused.
The block type selector: email, domain, disposable email, Discord ID, IP address, IP range, country, city, ISP, ASN, VPN, Tor, hosting, user agent, device fingerprint and card fingerprint

Sixteen block types, in full

  • Email
  • Domain
  • Disposable email
  • Discord ID
  • IP address
  • IP range
  • Country
  • City
  • ISP
  • ASN
  • VPN
  • Tor
  • Hosting
  • User agent
  • Device fingerprint
  • Card fingerprint

The four with a dot take no value. You switch those on and type nothing.

Dry run

Check a buyer before you block one.

Paste an email, an IP or a device fingerprint and Fraud Shield shows the verdict it would give at checkout. Nobody gets blocked, nothing is stored, and it never turns up in your event feed.

Allowlist first
Email, IP, IP range or device. One match exempts that buyer from every block you have.
Auto block
Pick the failed-payment count and the window. Fraud Shield then blocks the email, IP and device that order exposed.
On the invoice
The risk score and the signals behind it stay attached to the order, and they turn up again as evidence.
The Check a buyer dialog in the dashboard, with an email, an IP address and a device fingerprint filled in and a Run Check button

Chargeback Shield

Chargebacks run on a deadline. Chargeback Shield gets there first.

Most sellers never answer a dispute, so the bank hears one side and the money is gone. Chargeback Shield answers every one of yours with proof: Stripe or PayPal opens the case, and the file is built from your own delivery data while the dispute is still new. You read the email; the evidence is already stacked. It is on both plans, and there is no success fee.

The Disputes tab of Chargeback Shield, with the example case running from dispute opened through evidence collected and response submitted to the bank decision

What goes in the file

01Checkout device and network
The IP the order came from, its country, the user agent, and the device fingerprint captured at checkout.
02The fraud snapshot
The risk score Fraud Shield gave that order, and whether the connection was VPN, proxy or Tor at the time.
03Every access, timestamped
Downloads and delivery accesses, newest first, with IP and file name, as a fixed width table an issuer can read in a minute.
04Same IP as the payment
Chargeback Shield compares the access IP against the checkout IP and puts the answer in the statement.
05License activity
Activations, validation calls, last use, and whether the key is bound to the buyer hardware.
06What they bought before
Paid orders on that email before this one, the date of the first, and the total spent with you.
07Support contact
Tickets on the order or the email, with the last message date. Zero tickets is evidence too, and it says so.
08Terms and refund policy
Whether the buyer accepted them at checkout, with the timestamp they did it.

On top of those facts Chargeback Shield writes the cover statement the issuer reads first. It is slot filled from the same inventory, sentence by sentence, because no language model belongs on a path that moves money once and cannot take it back.

A missing fact stops the send
No verified download, no claim that you delivered. The case drops to your review naming the exact fields it lacks, and nothing half built goes to the bank. Six states run the case: new, evidence ready, submitted, then won, lost or accepted. The final submit fires once.
Visa Compelling Evidence 3.0
On Stripe fraud disputes, Chargeback Shield looks for two earlier paid orders from the same buyer, 120 to 364 days before this one, sharing at least two data elements with one of them a purchase IP, device fingerprint or device id. Stripe validates the qualification again, so nothing doomed gets staged. Nobody can promise you a win rate, and this page does not.
When it goes out
On autopilot Chargeback Shield files as soon as the packet is complete, and never inside the last two hours. If you approve cases yourself and have not, a sweep every 15 minutes files it 24 hours before the deadline. Under 48 hours with nothing submittable you get one alert, once.
Fraud warnings, before the dispute
The card networks flag some payments as fraud days before anyone charges back. Stripe forwards those, Chargeback Shield logs each one against the order, and you can have it refund on the spot while the amount is under your cap. That switch is yours and it ships off.

Ban Radar

0.75 percent is where Stripe starts watching.

Ban Radar counts your dispute rate every day and takes the worst number across your gateways, because your provider looks at each account on its own. Cross 0.4 percent and it warns you. Cross 0.75 and it can tighten checkout for you.

Protection levels and what each one turns on

Normal

Dispute rate under 0.4%

Checkout
Untouched
Card verification
Your own setting
Risk ceiling
Per product only
Purchases per buyer
Uncapped

Elevated

Dispute rate 0.4% and up

Checkout
Tightened
Card verification
3D Secure forced
Risk ceiling
Score 75 shop wide
Purchases per buyer
5 per IP, 5 per email, 24h

Critical

Dispute rate 0.75% and up

Checkout
Tightened
Card verification
3D Secure forced
Risk ceiling
Score 33 shop wide
Purchases per buyer
5, then 4, then 3 a day

Ban Radar watches from day one and moves nothing until you switch automatic protection on. Below 100 settled card sales it stays on monitoring, so a slow week cannot push you up a level. A first escalation is announced 24 hours ahead: apply it early, dismiss it, or pin the level where it is. Coming back down takes 14 clean days.

Checkout protection, three settings

OffDefault

Monitoring only. Ban Radar warns you and changes nothing about checkout.

Step-up

3D Secure on card payments, and VPN, proxy and high risk connections are refused.

Block

Everything step-up does, and the repeat purchase caps start rejecting orders.

Every tightening costs a little conversion, so the switch stays yours. Ban Radar ships watching and not touching: monitoring on, automatic protection off, checkout protection off.

Watch the number that gets shops closed.

Providers cut merchants off over a dispute rate, not over bad luck. Ban Radar counts it daily, and the other two keep it down.