Skip to content
TheChatArchiveThe Chat Archive
o4-mini-2025-04-16· OpenAIHugging Face LMArena Human Preference 140K

in js, how can i take only a domain and its tld in an arbitrary string that can contains…

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with gpt-4.1-mini-2025-04-14’s answer.
  1. User

    in js, how can i take only a domain and its tld in an arbitrary string that can contains anything, like query strings with domains, or sub-domains ?

  2. Assistant

    Here are two common ways to pull out just the “registered” domain + TLD (for example, from
    "Visit https://sub.foo.example.co.uk/path?x=1"
    you want "example.co.uk"):

    1. Use a Public Suffix List parser (most reliable)
    2. Use a quick & dirty regex + URL approach (no extra deps, but less foolproof)

    1) Using the “psl” (Public Suffix List) package

    The psl module knows about multi-level TLDs (e.g. “.co.uk”).

    npm install psl
    
    const psl = require('psl');
    
    function extractDomain(str) {
      // Find anything that looks like a hostname (with or without protocol)
      // then parse out the registered domain.
      const hostMatch = str.match(
        /([a-z0-9\-]+\.)+[a-z]{2,}(?=[\/:\s]|$)/gi
      );
      if (!hostMatch) return null;
    
      // We might get multiple hosts in the string; pick the first
      const host = hostMatch[0];
    
      // psl.get(...) returns e.g. "example.co.uk", or null if it’s an IP or invalid
      return psl.get(host);
    }
    
    // Examples
    console.log(extractDomain("Visit https://sub.foo.example.co.uk/path"));  
    // → "example.co.uk"
    
    console.log(extractDomain("Check example.com?x=1"));  
    // → "example.com"
    
    console.log(extractDomain("no-domain-here"));  
    // → null
    

    Pros: correctly handles foo.bar.example.co.uk → example.co.uk
    Cons: adds a small dependency (~200 KB unminified)


    2) Quick & Dirty with Regex + URL

    If you just need the last two labels (e.g. "example.com" from anything), you can:

    function extractDomainQuick(str) {
      // 1) Beef it up to a valid URL so URL parser will accept it
      const guessed =
        str.match(/https?:\/\//i) ? str : 'http://' + str.trim();
    
      let hostname;
      try {
        hostname = new URL(guessed).hostname;
      } catch {
        // no valid host found
        return null;
      }
    
      // 2) Split on dots, take the last two parts
      const parts = hostname.split('.');
      if (parts.length < 2) return null;
      const tld  = parts.pop();
      const dom  = parts.pop();
      return `${dom}.${tld}`;
    }
    
    // Examples
    console.log(extractDomainQuick("random text sub.foo.example.com/page"));  
    // → "example.com"
    
    console.log(extractDomainQuick("sub.fancy.site.co.uk"));  
    // → "co.uk"   👉 WRONG! you probably wanted "site.co.uk"
    

    Pros: zero dependencies, easy to drop in
    Cons: won’t handle multi-level TLDs correctly (e.g. .co.uk, .gov.au), and may pick up junk if your string contains weird dots


    Which to choose?

    • If you care about real registered domains (including .co.uk, .org.au, etc.), go with psl (or tldts / tldjs – same idea).
    • If you just need a quick hack for “always take the last two labels,” use the regex+URL trick.

    Let me know if you need examples for browser-only usage or a pure-regex-only version! 😊

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.