How to trace a spam page’s redirects with network tools

Spam messages often lead to a fake security screen that asks you to confirm you are human, press a button, or wait for a loading spinner. The page may then quietly send you elsewhere. This guide explains how to use network tools to see where a spam page is trying to redirect you without exposing your accounts or personal files. Learn more about Dating.

The useful distinction is between the address you entered and every address reached afterwards. A redirect can occur in an HTTP response, a script, a meta refresh, or only after the server matches your location, device, or referrer. The browser tab may show one destination while the network layer reveals several intermediate stops.

Start with developer tools in Chrome, Edge, or Firefox, then use command-line tools for more control. Treat every suspicious address as hostile: work in a disposable virtual machine or isolated browser profile, avoid entering credentials, and remove personal tokens from copied links. This is especially useful in Australia, where a plausible redirect can imitate a bank, telco, parcel service, or government campaign.

What a spam redirect can conceal

The first hop is rarely the whole story. A server may answer with 301 Moved Permanently, 302 Found, 307 Temporary Redirect, or 308 Permanent Redirect, placing the next address in the Location header. Each hop can add a tracking parameter, a new domain, or delay before the chain reaches a credential-harvesting page.

Verification screens can use less visible methods. HTML may contain a timed meta refresh; JavaScript may set window.location; an invisible frame may request another URL; or a server may vary its response according to IP address, user agent, language, cookies, and referral source. The same link can behave differently on Australian mobile data than on office broadband.

A trace shows where the browser went at one moment under one set of conditions. It does not certify the destination. A redirect through a familiar domain can still end at a fraudulent page, so record the whole route rather than judging by the final hostname.

Capture the requests in your browser

Open developer tools, select Network, enable Preserve log, and switch off the cache before reloading the suspicious page. Chrome and Edge usually open tools with F12; Firefox does too, although its filters differ. Do not click “Confirm”, “Continue”, or anything resembling a CAPTCHA during the test.

Filter for document requests first, then watch Fetch/XHR entries. A document row represents navigation, while fetch and XHR rows can reveal background requests that change the location later. Select a row without opening it elsewhere, then inspect its headers, timing, initiator, and any Location value. The Initiator panel often identifies the script or element responsible.

Use a fresh isolated profile if cookies might affect the result. Keep Preserve log active because a redirect creates another request that can otherwise vanish from view. “Copy as cURL” can move a request to a terminal, but remove cookies, authorisation headers, email addresses, and tracking identifiers before saving or sharing it.

Read status codes and Location headers

A 3xx response says the server has not delivered the final page. Codes 301 and 308 are cached more aggressively, while 302, 303, and 307 commonly send the client onward. A header such as Location: https://destination.example/path identifies the next hop and should be compared with the browser’s final URL.

The location may be absolute or relative. A path such as /go?id=42 must be joined to the current origin, and encoded characters may hide the real domain. A fragment after # is not normally sent to the server, so a script can add one without changing the server-side route.

Conditional redirects deserve special attention. A server might show a warning to one IP range and imitate a login to another, or use a cookie from an earlier visit. Repeat the request from the same clean profile, preserve the headers, and note the response size and content type. If the route changes, compare those conditions rather than assuming the first result applies everywhere.

Inspect the chain from a terminal

curl shows headers without opening a graphical browser. Run it in a disposable environment and start with a placeholder rather than clicking a live link. Replace it with the exact suspicious address, keeping the scheme and path but removing any personal token from the query string.

url='https://replace-with-the-untrusted-address'
curl -sS -D - -o /dev/null --max-redirs 0 --max-time 15 "$url"

This prints the initial response headers and refuses to follow a redirect, exposing the raw Location value. To inspect a bounded chain, add -L:

curl -sS -L -D - -o /dev/null --max-redirs 10 --max-time 30 "$url"

Review each status line and location in order. The limits stop an endless redirect loop from tying up the terminal. Curl requests can differ from browser requests because their user agent, cookies, language, and network path differ; that difference is evidence, not a defect. A sanitised browser export is closer to what the visitor actually received.

Separate DNS resolution from HTTP redirects

DNS answers identify the host that handles a name; they do not reveal every web address a page opens. dig +short example.com returns address records, while nslookup example.com offers a similar check. Repeat the lookup through your normal resolver and, if needed, a public resolver to expose split-horizon answers or cached addresses that differ by network.

Look for a CNAME chain as well. A spam page may use a CDN, parking service, or tracking domain before its HTTP response points elsewhere. Compare the returned address with the later Host header. Reverse DNS, WHOIS, and geolocation are clues rather than proof of ownership. An Australian .com.au or .au name can still be malicious, just as a .com name can be legitimate.

Use dig to test the name, not open the page. Avoid dig +trace on an untrusted domain unless you understand the extra DNS traffic. Resolver results can differ between Sydney and Melbourne, explaining why one carrier reaches a warning screen while another reaches the intended destination. Record the discrepancy because it may indicate regional targeting or CDN configuration.

Compare Australian network conditions

Australian users often move between fixed broadband and mobile data during commutes around Sydney, Melbourne, Brisbane, and Perth. Telstra, Optus, and Vodafone traffic can pass through different infrastructure, and an IP-based redirect may vary by state or network. A second isolated device on another connection can provide a useful comparison without asking anyone to interact with the page.

Local content can make a redirect convincing. Spam may imitate an Australian bank, mobile carrier, Australia Post delivery, Centrelink, or a state government notice, then use local urgency or deadline language. The informal phrase “that looks a bit dodgy” is not a technical control. Check the actual domain, certificate, redirect chain, and brand spelling; familiar logos and language can be copied easily.

Logs and capture tools often use UTC, while Australian readers may use AEST or AEDT depending on the state and season. Record UTC and add local time. Requests can also vary through Accept-Language, an en-AU preference, referrer data, or device fingerprinting. In a quarantined report, an analyst might catalogue a destination such as redirect destination examples, but it should remain hostile evidence rather than a live link on a personal device.

Use a safe inspection checklist

Before testing, prepare a clean environment and a reliable record. Keep the original message, because its sender address, embedded link, and received time may be more valuable than the destination alone.

After testing, preserve a compact evidence bundle rather than unrelated screenshots. Include the initial request and later navigation requests so a conditional redirect can be understood.

Interpretation comes last. If a second request produces no redirect, compare cookies, IP address, language, and referrer with the first run; the absence of a visible jump does not clear the domain. If credentials or money may have been exposed, changing the relevant password and contacting the bank promptly matters more than completing every technical check. The practical takeaway is to preserve the smallest useful evidence bundle—the first response, complete redirect chain, DNS answer, and final destination—while keeping every test away from accounts that matter.