<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>HeurChain blog</title>
    <link>https://heurchain.com/blog</link>
    <description>Notes on persistent memory, agent context, and building reliable AI workflows.</description>
    <language>en</language>
    <atom:link href="https://heurchain.com/blog/feed.xml" rel="self" type="application/rss+xml" />
    <lastBuildDate>Wed, 13 May 2026 00:00:00 GMT</lastBuildDate>
    <item>
      <title>Four types of agent amnesia (and the fix)</title>
      <link>https://heurchain.com/blog/four-types-of-agent-amnesia</link>
      <guid isPermaLink="true">https://heurchain.com/blog/four-types-of-agent-amnesia</guid>
      <pubDate>Wed, 13 May 2026 00:00:00 GMT</pubDate>
      <description>A coding agent could not find GitHub credentials sitting in local storage. Four kinds of agent amnesia, and the memory primitives that fix them.</description>
    </item>
    <item>
      <title>Introducing HeurChain</title>
      <link>https://heurchain.com/blog/introducing-heurchain</link>
      <guid isPermaLink="true">https://heurchain.com/blog/introducing-heurchain</guid>
      <pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate>
      <description>We built HeurChain because every AI tool we used had its own memory — isolated, ephemeral, and invisible to every other tool. Here is what we built and why.</description>
    </item>
    <item>
      <title>What happens when your AI runs out of context</title>
      <link>https://heurchain.com/blog/what-happens-when-your-ai-runs-out-of-context</link>
      <guid isPermaLink="true">https://heurchain.com/blog/what-happens-when-your-ai-runs-out-of-context</guid>
      <pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate>
      <description>An AI agent explains what happens to a conversation when the context buffer fills, and why nothing survives unless you explicitly save it.</description>
    </item>
    <item>
      <title>What persistent memory actually means for agents</title>
      <link>https://heurchain.com/blog/what-persistent-memory-actually-means-for-agents</link>
      <guid isPermaLink="true">https://heurchain.com/blog/what-persistent-memory-actually-means-for-agents</guid>
      <pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate>
      <description>Persistent memory is used loosely to mean anything from a long context window to a vector database. The differences matter for what you can actually build.</description>
    </item>
    <item>
      <title>The 512-token chunk pattern</title>
      <link>https://heurchain.com/blog/the-512-token-chunk-pattern</link>
      <guid isPermaLink="true">https://heurchain.com/blog/the-512-token-chunk-pattern</guid>
      <pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate>
      <description>Why we chunk agent memory at 512 tokens, how the bge-base-en-v1.5 context window shapes that decision, and what to do when your content does not fit neatly.</description>
    </item>
    <item>
      <title>Vault export and why data sovereignty matters</title>
      <link>https://heurchain.com/blog/vault-export-and-why-data-sovereignty-matters</link>
      <guid isPermaLink="true">https://heurchain.com/blog/vault-export-and-why-data-sovereignty-matters</guid>
      <pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate>
      <description>One-click vault download exists because the alternative, your AI memory trapped inside a vendor, is a structural problem for long-term AI use.</description>
    </item>
    <item>
      <title>Comparing HeurChain to Pinecone for agent memory</title>
      <link>https://heurchain.com/blog/comparing-heurchain-to-pinecone-for-agent-memory</link>
      <guid isPermaLink="true">https://heurchain.com/blog/comparing-heurchain-to-pinecone-for-agent-memory</guid>
      <pubDate>Tue, 12 May 2026 00:00:00 GMT</pubDate>
      <description>Honest comparison of HeurChain against Pinecone, Mem.ai, and Cognee for agent memory use cases — where each wins and where each falls short.</description>
    </item>
  </channel>
</rss>
