<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>ClawPipe Blog &amp; Changelog</title>
    <link>https://clawpipe.ai/blog/</link>
    <description>Engineering deep-dives, release notes, and customer stories from ClawPipe — the only AI gateway that skips LLM calls.</description>
    <language>en-us</language>
    <atom:link href="https://clawpipe.ai/blog/feed.xml" rel="self" type="application/rss+xml"/>
    <lastBuildDate>Wed, 30 Apr 2026 00:00:00 GMT</lastBuildDate>
    <generator>ClawPipe</generator>
    <image>
      <url>https://clawpipe.ai/og-cover.svg</url>
      <title>ClawPipe</title>
      <link>https://clawpipe.ai/blog/</link>
    </image>

    <item>
      <title>v3.7.0 — Provider failover, webhook DLQ, OpenAPI self-discovery, 90%+ coverage</title>
      <link>https://github.com/finsavvyai/clawpipe-sdk/releases/tag/v3.7.0</link>
      <guid isPermaLink="false">clawpipe-changelog-3.7.0</guid>
      <pubDate>Wed, 30 Apr 2026 00:00:00 GMT</pubDate>
      <description>SDK gains a provider failover chain with per-router health tracking. Gateway gains 11 features: Idempotency-Key, RFC 9239 RateLimit headers, W3C Trace Context, Webhook DLQ + retry, SSE Last-Event-ID resume, self-served OpenAPI 3.1, public stats endpoint, per-project savings, free-tier attribution, D1 read replication, KV cache warming. Test coverage on the gateway moves from 39% line / 74% branch to 90.15% line / 85.85% branch — both /CLAUDE.md gates now pass. Vitest bumped to 3.2.4 across all packages, clearing all dev-dep audit findings.</description>
    </item>

    <item>
      <title>How to use ClawPipe — 5-minute setup</title>
      <link>https://clawpipe.ai/blog/how-to-use-clawpipe</link>
      <guid isPermaLink="true">https://clawpipe.ai/blog/how-to-use-clawpipe</guid>
      <pubDate>Sun, 19 Apr 2026 00:00:00 GMT</pubDate>
      <description>Install the SDK, generate an API key, route your first prompt. Includes the OpenAI-compat shim for drop-in migrations from existing OpenAI client code.</description>
    </item>

  </channel>
</rss>
