<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
  <channel>
    <title>DEV Community: Joe Bromley</title>
    <description>The latest articles on DEV Community by Joe Bromley (@joe_bromley_934b9d223186c).</description>
    <link>https://dev.to/joe_bromley_934b9d223186c</link>
    <image>
      <url>https://media2.dev.to/dynamic/image/width=90,height=90,fit=cover,gravity=auto,format=auto/https:%2F%2Fdev-to-uploads.s3.us-east-2.amazonaws.com%2Fuploads%2Fuser%2Fprofile_image%2F3988963%2F6731c93d-1a64-479a-b55a-4f901dab7d44.png</url>
      <title>DEV Community: Joe Bromley</title>
      <link>https://dev.to/joe_bromley_934b9d223186c</link>
    </image>
    <atom:link rel="self" type="application/rss+xml" href="https://dev.to/feed/joe_bromley_934b9d223186c"/>
    <language>en</language>
    <item>
      <title>I built a free SEO content-cluster generator on Google Autocomplete data (no LLM, no API key)</title>
      <dc:creator>Joe Bromley</dc:creator>
      <pubDate>Wed, 17 Jun 2026 13:53:16 +0000</pubDate>
      <link>https://dev.to/joe_bromley_934b9d223186c/i-built-a-free-seo-content-cluster-generator-on-google-autocomplete-data-no-llm-no-api-key-5408</link>
      <guid>https://dev.to/joe_bromley_934b9d223186c/i-built-a-free-seo-content-cluster-generator-on-google-autocomplete-data-no-llm-no-api-key-5408</guid>
      <description>&lt;p&gt;kept hitting the same wall doing SEO: every tool hands you a keyword list, but none hand you an actual content &lt;em&gt;map&lt;/em&gt; you can plan a whole site around.&lt;/p&gt;

&lt;p&gt;So I built &lt;a href="https://topicmap.app" rel="noopener noreferrer"&gt;TopicMap&lt;/a&gt; - paste one topic, get 60-90 article ideas grouped into 7-9 pillars, each tagged with search intent. A cluster you can hand straight to a writer. Free, no login.&lt;/p&gt;

&lt;h2&gt;
  
  
  The interesting part: no LLM
&lt;/h2&gt;

&lt;p&gt;The first version used an LLM to generate the map. Two problems: it cost money per generation, and the difficulty scores were guesses - a real SEO would never trust them.&lt;/p&gt;

&lt;p&gt;So I rebuilt it on real Google Autocomplete data instead:&lt;/p&gt;

&lt;ul&gt;
&lt;li&gt;Recursively expands your topic through Google's Suggest endpoint (2 levels deep) to harvest 100-800 real queries people actually type&lt;/li&gt;
&lt;li&gt;Clusters them into pillars by intent (how-to / commercial / transactional / questions)&lt;/li&gt;
&lt;li&gt;Filters out local-business + geo noise&lt;/li&gt;
&lt;li&gt;Tags each with intent + an estimated difficulty&lt;/li&gt;
&lt;/ul&gt;

&lt;p&gt;No API key, no per-call cost, ~4s per map. Every keyword is a query people genuinely search - not invented.&lt;/p&gt;

&lt;h2&gt;
  
  
  Honest caveats
&lt;/h2&gt;

&lt;ul&gt;
&lt;li&gt;Difficulty is &lt;em&gt;estimated&lt;/em&gt; (from query specificity), not live SERP data. That's the next thing if people find it useful.&lt;/li&gt;
&lt;li&gt;Narrow topics return thinner maps - because there's genuinely less real search demand. It shows you what's real, it doesn't pad.&lt;/li&gt;
&lt;/ul&gt;

&lt;h2&gt;
  
  
  Try it
&lt;/h2&gt;

&lt;p&gt;Free, no login: &lt;a href="https://topicmap.app" rel="noopener noreferrer"&gt;https://topicmap.app&lt;/a&gt;&lt;/p&gt;

&lt;p&gt;Would love blunt feedback from SEO + content folks: is the output good enough to plan real content from, or does it read thin? What would make it worth paying for?&lt;/p&gt;



&lt;div class="highlight js-code-highlight"&gt;
&lt;pre class="highlight plaintext"&gt;&lt;code&gt;

&lt;/code&gt;&lt;/pre&gt;

&lt;/div&gt;

</description>
      <category>seo</category>
      <category>webdev</category>
      <category>buildinpublic</category>
      <category>startup</category>
    </item>
  </channel>
</rss>
