<?xml version="1.0" encoding="UTF-8"?>
  <rss version="2.0">
    <channel>
      <title>Threada Blog</title>
  <link>https://threada.ai/blog/</link>
      <description>AI-native shells, packs, flagship apps, retrieval engineering, and evaluation operations.</description>
      
    <item>
      <title><![CDATA[Automated Resolution: The Honest Meter]]></title>
      <link>https://threada.ai/blog/automated-resolution-the-honest-meter/</link>
      <guid>https://threada.ai/blog/automated-resolution-the-honest-meter/</guid>
      <pubDate>Fri, 05 Jun 2026 00:00:00 GMT</pubDate>
      <description><![CDATA[Why Threada's enterprise pricing unit is the automated resolution — work completed without human takeover — and why metering outcomes is more honest than metering messages or seats.]]></description>
    </item>

    <item>
      <title><![CDATA[The Five Surfaces of Governed Work]]></title>
      <link>https://threada.ai/blog/five-surfaces-of-governed-work/</link>
      <guid>https://threada.ai/blog/five-surfaces-of-governed-work/</guid>
      <pubDate>Fri, 05 Jun 2026 00:00:00 GMT</pubDate>
      <description><![CDATA[Threada decomposes a workspace into five surfaces — intent, canvas, evidence, controls, and run log. Here is what each one is for and why the split matters.]]></description>
    </item>

    <item>
      <title><![CDATA[Scenario: Running Vendor Security Reviews Through Threada]]></title>
      <link>https://threada.ai/blog/scenario-vendor-security-review/</link>
      <guid>https://threada.ai/blog/scenario-vendor-security-review/</guid>
      <pubDate>Fri, 05 Jun 2026 00:00:00 GMT</pubDate>
      <description><![CDATA[An illustrative walkthrough — not a customer story — of how a security team would run a vendor review through Threada's governed workflow, from intake to a recorded decision.]]></description>
    </item>

    <item>
      <title><![CDATA[Why Every Action Should Leave a Receipt]]></title>
      <link>https://threada.ai/blog/why-every-action-leaves-a-receipt/</link>
      <guid>https://threada.ai/blog/why-every-action-leaves-a-receipt/</guid>
      <pubDate>Fri, 05 Jun 2026 00:00:00 GMT</pubDate>
      <description><![CDATA[Threada's governance model treats auditability as a default, not an add-on. Here is why every governed action produces a receipt, and what that buys an operations team.]]></description>
    </item>

    <item>
      <title><![CDATA[Chunking Strategies for RAG]]></title>
      <link>https://threada.ai/blog/chunking-strategies/</link>
      <guid>https://threada.ai/blog/chunking-strategies/</guid>
      <pubDate>Mon, 15 Sep 2025 00:00:00 GMT</pubDate>
      <description><![CDATA[Designing high-precision chunking for website RAG: fixed, semantic, hierarchical, adaptive strategies and evaluation.]]></description>
    </item>

    <item>
      <title><![CDATA[Hybrid Retrieval: Vector + Keyword + Metadata]]></title>
      <link>https://threada.ai/blog/hybrid-retrieval/</link>
      <guid>https://threada.ai/blog/hybrid-retrieval/</guid>
      <pubDate>Mon, 15 Sep 2025 00:00:00 GMT</pubDate>
      <description><![CDATA[Engineering resilient hybrid retrieval: blending vector, lexical, metadata & temporal signals for website RAG.]]></description>
    </item>

    <item>
      <title><![CDATA[Prompt Versioning and Rollback Strategy]]></title>
      <link>https://threada.ai/blog/prompt-versioning-and-rollback-strategy/</link>
      <guid>https://threada.ai/blog/prompt-versioning-and-rollback-strategy/</guid>
      <pubDate>Mon, 24 Feb 2025 00:00:00 GMT</pubDate>
      <description><![CDATA[Treat prompts like code: version them, test in staging, launch with annotations, and roll back quickly if metrics slip.]]></description>
    </item>
    </channel>
  </rss>