<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
<channel>
  <title>Atelier Blog</title>
  <link>https://atelier.ws/blog</link>
  <description>Engineering notes about coding-agent runtimes, code intelligence, context engineering, and cost.</description>
  <language>en</language>
  <lastBuildDate>Tue, 23 Jun 2026 00:00:00 GMT</lastBuildDate>
  <atom:link href="https://atelier.ws/rss.xml" rel="self" type="application/rss+xml" />
  <item>
    <title>Context Engineering: What Atelier Shows the Agent</title>
    <link>https://atelier.ws/blog/context-engineering-in-atelier</link>
    <guid isPermaLink="true">https://atelier.ws/blog/context-engineering-in-atelier</guid>
    <description>A coding agent is only as good as the context it carries. Atelier treats that context as an engineering problem: progressive disclosure, bounded results, deduplication, budgeted selection, cache-stable ordering, and boundary-aware compaction.</description>
    <author>contact@atelier.ws (Atelier)</author>
    <pubDate>Tue, 23 Jun 2026 00:00:00 GMT</pubDate>
    <category>Context engineering</category>
    <category>Code intelligence</category>
    <category>Coding agents</category>
  </item>
  <item>
    <title>Five Kinds of Memory: How Atelier Remembers</title>
    <link>https://atelier.ws/blog/five-kinds-of-memory-in-atelier</link>
    <guid isPermaLink="true">https://atelier.ws/blog/five-kinds-of-memory-in-atelier</guid>
    <description>Memory is the most overloaded word in agent tooling. Atelier uses it for five genuinely different systems: curated facts, searchable history, write-time arbitration, read-only cross-vendor ingestion, and a content-addressed cache of code structure.</description>
    <author>contact@atelier.ws (Atelier)</author>
    <pubDate>Tue, 23 Jun 2026 00:00:00 GMT</pubDate>
    <category>Memory</category>
    <category>Code intelligence</category>
    <category>Coding agents</category>
  </item>
  <item>
    <title>How Atelier Works: MCP Runtime for Coding Agents</title>
    <link>https://atelier.ws/blog/how-atelier-works</link>
    <guid isPermaLink="true">https://atelier.ws/blog/how-atelier-works</guid>
    <description>Atelier is an MCP runtime between your coding agent and your repository. It changes how the agent gathers context, makes edits, and carries work across a long session.</description>
    <author>contact@atelier.ws (Atelier)</author>
    <pubDate>Mon, 22 Jun 2026 00:00:00 GMT</pubDate>
    <category>MCP</category>
    <category>Code intelligence</category>
    <category>Coding agents</category>
  </item>
  <item>
    <title>How Atelier Saves Money on Coding-Agent Tokens</title>
    <link>https://atelier.ws/blog/how-atelier-saves-money</link>
    <guid isPermaLink="true">https://atelier.ws/blog/how-atelier-saves-money</guid>
    <description>Atelier saves money by changing what enters the model&apos;s context: fewer discovery turns, smaller source views, bounded results, and less repeated text.</description>
    <author>contact@atelier.ws (Atelier)</author>
    <pubDate>Sun, 21 Jun 2026 00:00:00 GMT</pubDate>
    <category>Cost</category>
    <category>Context engineering</category>
    <category>Coding agents</category>
  </item>
  <item>
    <title>Inside Atelier: technology and design ideas</title>
    <link>https://atelier.ws/blog/inside-atelier-technology-and-ideas</link>
    <guid isPermaLink="true">https://atelier.ws/blog/inside-atelier-technology-and-ideas</guid>
    <description>An inventory of Atelier&apos;s public tools, internal capabilities, technology stack, and the engineering ideas that connect them.</description>
    <author>contact@atelier.ws (Atelier)</author>
    <pubDate>Sat, 20 Jun 2026 00:00:00 GMT</pubDate>
    <category>Architecture</category>
    <category>Technology</category>
    <category>MCP</category>
  </item>
</channel>
</rss>
