← All guides
Proxy & VPN detection

RedTrack fraud filter: score and block proxy, VPN and bot traffic

RedTrack filters proxy types and blacklists bots, but has no external scoring hook. Add a FraudNova redirect gate as your lander and block by score.

Published

RedTrack is a cloud tracker: in its own words, "Cloud. No server, no VPS, no sysadmin." That is convenient, but it also means you cannot install anything inside it. This guide adds a FraudNova score to RedTrack campaigns anyway, by putting a small redirect gate of your own in the funnel as the lander, so risky visitors meet a safe page before they ever see your offer.

The fraud filtering for affiliate trackers overview explains the gate in depth and compares RedTrack with Keitaro, Binom and Voluum. Here we stick to the RedTrack specifics.

What RedTrack filters on its own

RedTrack's traffic filters cover a long list: Countries, Regions, Cities, ISP, Browser, OS, device brand, model and type, IP, Sub-values, Languages, Referrer, Connection types, Unique path, and two that sound like fraud protection:

  • Proxy types, with three choices: Anonymous, Transparent and Hosting.
  • Fraud check, which appears in the filter list, but the help pages do not say what it checks or where its data comes from. Test it before you rely on it.

Traffic that fails a filter goes to the campaign's Fallback URL, or you can send it to another funnel, lander or offer inside the same campaign.

Separately, Tools → Blacklist bots blocks IP ranges and user agents you list. You find those bots yourself (RedTrack also offers a hidden "bot trap" link to catch crawlers), and blocked traffic is left out of your reports.

Two limits shape everything below. First, RedTrack's help is clear that filtering needs redirect tracking: "With no-redirect tracking, filtering works on the offer level only." Blacklist bots is also redirect-only. Second, RedTrack documents no custom code and no per-click API hook. You cannot ask it to call a scoring service and branch on the reply. So none of the native options gives you a per-visitor risk score, and that is the part FraudNova supplies.

What the gate does for a RedTrack click

The gate is open source; the code lives in our examples repository under examples/tracker-gate. Pick the PHP version (gate.php, PHP 8.2 with Composer) or the Node.js one, and run it on any small host you control. RedTrack itself stays untouched.

For each visit it checks the IP the visitor actually connects from, sending the browser's User-Agent and Accept-Language along. Nothing in the URL can override that IP. The answer is a fraud_score from 0 to 100 plus true/false flags including proxy, vpn, tor and bot_status, all described on the proxy and VPN detection API page.

By default a click is blocked at a score of 75 or more, or when any of those four flags is true. active_vpn can be switched on as an extra flag, and allowed_countries limits traffic to the markets you buy. Blocked visitors get a 302 to your safe URL, the rest a 302 to your money URL, and the query string comes with them, so RedTrack's clickid is still on the URL afterwards. If the check fails or takes longer than 2 seconds, the click goes through, because a slow API should never cost you a sale. You can make it fail closed in the config if you prefer.

Redirect or no-redirect: check this first

RedTrack offers both tracking modes on every plan, but a campaign using the gate must use redirect tracking. If it runs on the no-redirect script, the visitor goes straight to your page and never passes through a lander URL you control, so the gate cannot sit in front of it. Switch that campaign to redirect tracking first, or keep it on RedTrack's own offer-level filters.

Set up the gate in RedTrack

Deploy the gate and edit its config: your FraudNova API key (fn_..., from Dashboard, API keys) and a default route whose money URL is your real landing page and whose safe URL is a neutral page.

Then:

  1. Under Landers → New, add a lander whose URL is the gate:

    https://gate.example.com/gate.php?clickid={clickid}&route=default
    
  2. In the campaign, build the funnel as Lander → Offer, with this gate lander in the lander slot and your offer as before.

  3. Leave the call-to-action on your real landing page as it is. It must still be RedTrack's LP click URL, https://your.tracking.domain/click (or /click/1, /click/2, /click/3 when the page carries several offers). RedTrack calls that "obligatory for all campaigns with landing pages", and it keeps the path from your page to the offer running through RedTrack.

RedTrack lists {clickid} among its general tokens and says they can be used "for a traffic channel, affiliate network, or a Lander", but its help shows no lander URL example. Send one test click and, if you enabled the gate's log_file, look at its log, or check the address bar: you should see a real click id, not the literal text {clickid}. Do not add {ip} or {useragent} to the URL; the gate reads those from the visitor directly, which cannot be faked.

Send blocked clicks to your RedTrack reports

Because a blocked visitor stops at the gate, RedTrack would only see a click with no lander-to-offer step. To label those clicks, record them as a custom conversion type.

  1. Go to Tools → Conversion tracking → Conversion types and create a type named blocked. RedTrack warns that "The fields are case-sensitive", so the name must match the URL exactly.

  2. Take the S2S postback URL RedTrack generated when you saved your offer source, set its click id to {clickid}, append &type=blocked, and leave out the payout value, since a blocked click earns nothing. It will look roughly like this:

    https://your.tracking.domain/postback?clickid={clickid}&type=blocked
    
  3. Put that URL in the gate's report_url.

The gate requests it once for each blocked click and never for allowed ones. Custom types only show up in reports after you add them as columns, so add a blocked column to the campaign report. Statuses (approved, pending, declined, other) are fixed in RedTrack and cannot be renamed, which is why a conversion type is the right tool here.

Picking a threshold for RedTrack traffic

FraudNova scores fall into three bands: 0–39 clean, 40–74 suspicious, 75–100 high risk. The gate's default of 75 blocks only the high-risk band. On lead offers where a rejected lead hurts, try 60 or so; on cheap pops or push traffic you might stay at 75 and rely on the flags. Change one thing at a time and compare the blocked column against conversions. The fraud score thresholds guide goes deeper.

If you also keep RedTrack's Proxy types filter on the campaign, the visitors it catches go to the Fallback URL and never cost a check. That saves credits, but you then have two judges with different opinions, so watch both.

Credits

A check costs 1 credit, charged the first time the gate sees an IP in a day; repeat clicks from that IP within 24 hours reuse the cached verdict for free. New accounts get 25 credits after verifying their email, enough to try one campaign, and more are prepaid packs with no subscription. Rate limits are listed in the API docs.

Try it

Take a visitor IP from a recent RedTrack click and run it through the free IP checker: that is the verdict the gate would act on. When the results look right, create an account, deploy the gate on one redirect campaign, and watch the blocked column for a week before rolling it out.

Try it free

Run one IP proxy check right now — no account, no card.

Try the free IP proxy checker — no signup See the Proxy & VPN detection API

Also read