Back to Blog
typosquattingphishingbrand protectionDNS

Typosquatting: How Attackers Use Fake Domains to Impersonate Your Brand

Attackers register domains that look nearly identical to yours — then use them for phishing, malware delivery, and brand fraud. Here's how to detect lookalike domains before they're weaponized.

ThreatPort Team
April 14, 2026
9 min read

BLUF: Lookalike Domains Are Cheap to Register and Hard to Catch

Typosquatting is the practice of registering domain names that are visually or phonetically similar to a legitimate brand, then weaponizing them for phishing, malware delivery, brand impersonation, or credential harvesting. Generating plausible typo variants of a single domain produces 40-100 candidates, and with over 1,500 active top-level domains (TLDs) available as of 2025, an attacker can register lookalikes under .com, .net, .co, .io, .ai, or dozens of other extensions for under $15 each. Well over a million unique phishing attacks are observed in a typical quarter, and the majority use deceptive lookalike domains rather than compromised legitimate infrastructure. BEC attacks cost organisations billions of dollars a year, and many originate from domains differing by a single character. Defensive registration of every variant is economically untenable for most organizations. The practical control is continuous monitoring: detecting when lookalike domains are registered, when they acquire MX records (a strong signal of imminent phishing activity), or when SSL certificates are issued for them. The earlier you spot a malicious lookalike, the cheaper and faster the takedown.

What Is Typosquatting?

Typosquatting is the practice of registering domain names that are visually or phonetically similar to a legitimate domain, then using those domains for malicious purposes.

Attackers use a range of techniques: character substitution (rn in place of m), homoglyphs (Unicode characters that look identical to ASCII), missing letters (threatpor.com instead of threatport.com), extra letters (threattport.com), different TLDs (yourcompany.net when yourcompany.com is legitimate), hyphens added or removed, and plural/singular swaps.

An attacker generating typo variants of a single domain can produce 40-100 plausible candidates. With hundreds of millions of registered domains and thousands of new TLDs launched since 2013, the combinatorial space for lookalike domains is vast.

Homograph and Punycode Attacks

A particularly deceptive variant is the homograph attack, which leverages Internationalized Domain Names (IDN) and Punycode to register domains that appear visually identical to the legitimate brand. For example, "threatport.com" can be registered using the Cyrillic "а" (Unicode U+0430) instead of the Latin "a", producing a domain that renders identically in browsers but resolves to an entirely different IP. Large numbers of internationalised domain registrations display strings that are visually confusable with popular brand domains.

The Unicode confusables list maintained by the Unicode Consortium documents hundreds of character pairs or groups that are visually similar across scripts. Attackers exploit these systematically: the Latin "o", Cyrillic "о", and Greek "ο" all render identically in most fonts, yet each is a different character point. A domain name containing one of these characters is technically distinct under ICANN rules, making detection reliant on visual inspection rather than automated string matching.

Modern browsers attempt to flag obvious homograph attacks, but coverage is inconsistent and users on mobile devices are especially vulnerable. The IDN homograph problem has never been solved at the protocol level. Browsers apply heuristics, and attackers learn to evade them.

What Attackers Do With Lookalike Domains

Phishing campaigns: Send emails from [email protected] with a realistic invoice. The "From" address looks legitimate at a glance, and recipients who do not scrutinize each character may not notice the extra "o". Lookalike domains are now the dominant vehicle for phishing campaigns, ahead of compromised legitimate sites, because registering a new domain is cheaper and faster than breaking into someone else's.

Business email compromise (BEC): Register yourcompany-invoicing.com and use it to impersonate your finance department when communicating with vendors or clients. BEC attackers often register these domains weeks in advance, letting them age slightly to bypass domain-age reputation checks.

Malware distribution: Host download pages that serve malware under your brand name. Common in software supply chain attacks where the attacker targets users searching for legitimate software downloads.

Credential harvesting: Create a pixel-perfect copy of your login page on the lookalike domain. These pages often include real-time credential forwarding to the legitimate site so the user sees no error and never realizes they were phished.

Traffic interception: Register the common typo of your URL and serve ads, competitor content, or malware to users who mistype. Some typosquatters monetize purely through ad revenue, generating passive income from traffic that was intended for your site.

The Detection Challenge

There are over 1,500 active top-level domains. For each variant of your name, an attacker can register it under any of them. You cannot defensively register all possible variations — there are too many and it would cost tens of thousands of dollars per year.

The practical approach is monitoring: continuously check which typo variants have active DNS records or mail exchangers (MX records). An MX record means someone has configured the domain to send or receive email — a strong signal of active malicious use. Certificate transparency logs provide another high-confidence signal: when an SSL certificate is issued for a lookalike domain, it indicates the registrant is actively preparing a phishing site or credential harvesting page.

Prioritizing What Matters

Not every registered lookalike is immediately dangerous. The critical signals are active MX records (the domain is configured to send email), active A/CNAME records pointing to hosting, and SSL certificates issued (someone is actively using it for HTTPS). A lookalike domain with none of these signals may be a defensive registration by a security researcher or simply an unmonetized squatter.

Most phishing registrations are used within days of the domain being created, so detection has to operate in hours rather than months to matter at all. The window between registration and weaponization is narrowing as attackers automate their infrastructure setup.

Brand Protection Techniques

  1. Defensive registration of the highest-risk typos and alternative TLDs (.com, .net, .org, .co). Limit this to the top 10-15 variants; cost-benefit drops sharply beyond that.
  1. Trademarks filed in your operating jurisdictions give you UDRP arbitration leverage for takedowns.
  1. Continuous monitoring of certificate transparency logs and passive DNS for lookalike domains matching your brand.
  1. DMARC p=reject to prevent spoofing of your own domain — forcing attackers to register lookalikes, which are easier to detect.

Takedowns

When you find an active malicious lookalike: document everything (screenshots, WHOIS data, DNS records, content being served), report to the registrar, report to hosting providers, consider UDRP arbitration for persistent trademark infringement, and report to anti-phishing organizations (Google Safe Browsing, Microsoft SmartScreen, PhishTank).

Takedowns take time. Quick-action channels through hosting providers can take a domain offline in hours, while UDRP arbitration typically takes 2-3 months. The monitoring is the early warning system that lets you act before a phishing campaign reaches your customers.

Frequently Asked Questions

Should I register all typo variants of my domain?

No. A major brand can generate hundreds of plausible typos across dozens of TLDs, and registering all of them would cost tens of thousands of dollars annually while still missing the most inventive variants. Best practice, per ICANN guidance, is to defensively register the 10-15 highest-risk variants (closest typos, most common TLDs like .com, .net, .co, .io) and then rely on continuous monitoring for the rest. Monitoring through passive DNS and certificate transparency catches the new registrations within hours and lets you pursue takedowns through UDRP or registrar complaints. Defensive registration is insurance; monitoring is detection.

Can I detect a homograph attack before it reaches my customers?

Partially. Browser vendors maintain lists of lookalike domains and some flag obvious homograph attacks, but coverage is inconsistent. The reliable approach is to use brand-monitoring services that compare newly registered domains against your brand using Levenshtein distance, Unicode normalization, and IDN homograph detection libraries. These services identify lookalikes within hours of registration. Certificate transparency log monitoring belongs in the baseline set of controls: a new SSL certificate issued for a lookalike domain is one of the earliest available signals that a phishing campaign is being prepared.

How fast do I need to respond to a malicious lookalike?

Within hours to days for high-risk variants that show active MX records or SSL certificates. A phishing domain typically becomes actively malicious within a day or two of registration, which is why detection has to run continuously rather than on a review cycle. The quicker you detect and initiate a takedown, the less time attackers have to harvest credentials or distribute malware. For persistent trademark infringement, the UDRP process can take 2-3 months, so supplementary steps like Google Safe Browsing reports and hosting provider complaints are essential to neutralize immediate threats while legal channels run their course.

What is the difference between typosquatting and domain hijacking?

Typosquatting involves registering new domains that look similar to yours; the attacker never touches your legitimate infrastructure. Domain hijacking involves taking over your actual domain by compromising your registrar account, DNS provider, or email-based transfer authorization. Typosquatting is a brand protection problem; hijacking is a credential and account control problem. Both warrant monitoring, but the defensive controls differ: typosquatting defense focuses on lookalike detection and takedown, while hijacking defense focuses on MFA at the registrar, registry lock, and monitoring for unauthorized DNS changes. Confusing the two is a common mistake that leads to investing in the wrong defense.

How does DMARC help with typosquatting?

DMARC p=reject makes it impossible for attackers to spoof your genuine domain, forcing them to register lookalikes instead. Lookalike domains are easier to detect because they have to be actively registered, configured with DNS records, and SSL certificates issued — each of which leaves a footprint in WHOIS, passive DNS, and certificate transparency logs. Without DMARC enforcement, attackers can spoof your exact domain with no DNS footprint at all. With enforcement, you surface a high-value detection signal: any domain sending mail "as your company" outside your authorized senders is either a misconfigured partner or an attacker. DMARC turns the invisible problem of spoofing into the searchable problem of lookalike registration.

How are attackers using AI to generate lookalike domains?

Generative AI and machine learning tools have been adopted by attackers to generate and prioritize lookalike domain candidates at scale. Instead of manually applying character substitution rules, attackers can train models on brand names, common user typos, and click-through data to predict which lookalike variants are most likely to fool users. AI can also automate the registration process, creating hundreds of lookalike domains across multiple TLDs within minutes. AI-generated typosquatting has grown sharply, and the detection problem now includes algorithmically optimised variants designed to slip past string-similarity filters, not just human-plausible typos. Defenders must respond by adopting AI-powered detection tools that analyze behavioral signals — DNS resolution patterns, certificate issuance timing, and email sending volume — rather than relying solely on static string matching.

We use essential cookies to operate this site and optional analytics cookies to improve it. Cookie Policy