<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>MDDB blog</title>
    <link>https://mddb.tradik.com/</link>
    <description>MDDB blog</description>
    <atom:link href="https://mddb.tradik.com/rss.xml" rel="self" type="application/rss+xml"/>
    <lastBuildDate>Mon, 24 Aug 2026 19:46:12 +0000</lastBuildDate>
    <item>
      <title>Is It Leaking? Read a Different Number</title>
      <link>https://mddb.tradik.com/blog/rss-is-the-wrong-number/</link>
      <guid isPermaLink="false">https://mddb.tradik.com/blog/rss-is-the-wrong-number/</guid>
      <pubDate>Mon, 24 Aug 2026 00:00:00 +0000</pubDate>
      <category>operations</category>
      <category>memory</category>
      <category>profiling</category>
      <category>monitoring</category>
      <description>A report said MDDB grew from 42 MB to 153 MB under load. The observation was probably right and the conclusion wrong — and the difference is one field you already have.</description>
    </item>
    <item>
      <title>Flat Throughput Is a Diagnosis</title>
      <link>https://mddb.tradik.com/blog/the-loop-that-was-not-waiting-for-memory/</link>
      <guid isPermaLink="false">https://mddb.tradik.com/blog/the-loop-that-was-not-waiting-for-memory/</guid>
      <pubDate>Mon, 24 Aug 2026 00:00:00 +0000</pubDate>
      <category>performance</category>
      <category>vector-search</category>
      <category>simd</category>
      <category>benchmarking</category>
      <description>Two optimisations to MDDB&#39;s vector loop predicted 33% and delivered 1.5%. The benchmark had been saying why for months, in a column nobody read as a diagnosis.</description>
    </item>
    <item>
      <title>What&#39;s New in MDDB 2.12</title>
      <link>https://mddb.tradik.com/blog/whats-new-in-mddb-2-12/</link>
      <guid isPermaLink="false">https://mddb.tradik.com/blog/whats-new-in-mddb-2-12/</guid>
      <pubDate>Mon, 24 Aug 2026 00:00:00 +0000</pubDate>
      <category>release</category>
      <category>search</category>
      <category>rag</category>
      <category>security</category>
      <category>performance</category>
      <description>MDDB 2.12 adds code-aware retrieval, collection search profiles, faster bulk indexing, safer defaults and clearer production diagnostics.</description>
    </item>
    <item>
      <title>What&#39;s New in MDDB 2.13</title>
      <link>https://mddb.tradik.com/blog/whats-new-in-mddb-2-13/</link>
      <guid isPermaLink="false">https://mddb.tradik.com/blog/whats-new-in-mddb-2-13/</guid>
      <pubDate>Mon, 24 Aug 2026 00:00:00 +0000</pubDate>
      <category>release</category>
      <category>windows</category>
      <category>performance</category>
      <category>security</category>
      <category>operations</category>
      <description>MDDB 2.13 runs its test suite on Windows, makes vector comparison 3.7–6.3× faster on amd64, and fixes two ways a restore could destroy the database it was restoring.</description>
    </item>
    <item>
      <title>The Collection Knows How to Be Searched</title>
      <link>https://mddb.tradik.com/blog/the-collection-knows-how-to-be-searched/</link>
      <guid isPermaLink="false">https://mddb.tradik.com/blog/the-collection-knows-how-to-be-searched/</guid>
      <pubDate>Sun, 23 Aug 2026 00:00:00 +0000</pubDate>
      <category>rag</category>
      <category>retrieval</category>
      <category>mcp</category>
      <category>configuration</category>
      <category>agents</category>
      <description>Retrieval settings and answer formatting used to live in every client. In MDDB 2.12 they live with the data, so a new consumer gets them right without being told.</description>
    </item>
    <item>
      <title>What Breaks If I Change This?</title>
      <link>https://mddb.tradik.com/blog/what-breaks-if-i-change-this/</link>
      <guid isPermaLink="false">https://mddb.tradik.com/blog/what-breaks-if-i-change-this/</guid>
      <pubDate>Sun, 23 Aug 2026 00:00:00 +0000</pubDate>
      <category>code-intelligence</category>
      <category>search</category>
      <category>refactoring</category>
      <category>css</category>
      <category>mcp</category>
      <description>MDDB 2.12 indexes what a code file declares, uses and imports, then answers the relational questions grep cannot: what depends on this selector, which pages load this script, what is nobody using any more.</description>
    </item>
    <item>
      <title>Choosing a Search Algorithm from the Collection</title>
      <link>https://mddb.tradik.com/blog/which-algorithm-should-i-use/</link>
      <guid isPermaLink="false">https://mddb.tradik.com/blog/which-algorithm-should-i-use/</guid>
      <pubDate>Sun, 23 Aug 2026 00:00:00 +0000</pubDate>
      <category>search</category>
      <category>rag</category>
      <category>agents</category>
      <category>mcp</category>
      <category>benchmarking</category>
      <description>MDDB now profiles a collection and recommends search settings from its document length, vocabulary, vector coverage and extracted symbols.</description>
    </item>
    <item>
      <title>Diversifying Near-Duplicate Search Results</title>
      <link>https://mddb.tradik.com/blog/your-vector-search-rewards-duplicates/</link>
      <guid isPermaLink="false">https://mddb.tradik.com/blog/your-vector-search-rewards-duplicates/</guid>
      <pubDate>Sun, 23 Aug 2026 00:00:00 +0000</pubDate>
      <category>search</category>
      <category>rag</category>
      <category>ranking</category>
      <category>embeddings</category>
      <category>deduplication</category>
      <description>A runbook copied for four environments can occupy four result slots. MDDB uses MinHash overlap to reduce that repetition after ranking.</description>
    </item>
    <item>
      <title>Your WordPress just became AI-searchable: MDDB Sync</title>
      <link>https://mddb.tradik.com/blog/wordpress-plugin/</link>
      <guid isPermaLink="false">https://mddb.tradik.com/blog/wordpress-plugin/</guid>
      <pubDate>Sun, 02 Aug 2026 00:00:00 +0000</pubDate>
      <description>There&#39;s a decade of knowledge sitting in your WordPress. Product announcements,
how-to guides, changelogs, that one post from 2019 that still answers half of
your support tickets. And when you plug an AI agent into your company, the
first thing it asks is, effectively: where&#39;s the knowledge?
The MDD</description>
    </item>
    <item>
      <title>MDDB 2.11.4: tenants, better RAG, vectors that fit</title>
      <link>https://mddb.tradik.com/blog/mddb-2-11-4/</link>
      <guid isPermaLink="false">https://mddb.tradik.com/blog/mddb-2-11-4/</guid>
      <pubDate>Fri, 31 Jul 2026 00:00:00 +0000</pubDate>
      <description>We spent the last cycle reading a lot of feedback — not just about MDDB, but
about what people using vector databases in general keep running into. A few
themes came up again and again: &amp;quot;I want to put multiple customers on one
instance without building isolation myself&amp;quot;, &amp;quot;my RAG answe</description>
    </item>
  </channel>
</rss>
