<?xml version="1.0" encoding="utf-8" standalone="yes"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/">
  <channel>
    <title>OpenAI Responses on Mengboy Tech Notes</title>
    <link>https://www.mfun.ink/en/tags/openai-responses/</link>
    <description>Recent content in OpenAI Responses on Mengboy Tech Notes</description>
    <generator>Hugo -- 0.156.0</generator>
    <language>en</language>
    <lastBuildDate>Fri, 27 Mar 2026 01:08:00 +0000</lastBuildDate>
    <atom:link href="https://www.mfun.ink/en/tags/openai-responses/index.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>OpenAI Responses Streaming in Production: Backpressure, Chunk Reassembly, and Timeout Budget</title>
      <link>https://www.mfun.ink/en/2026/03/27/openai-responses-streaming-backpressure-chunk-reassembly-timeout-budget/</link>
      <pubDate>Fri, 27 Mar 2026 01:08:00 +0000</pubDate>
      <guid>https://www.mfun.ink/en/2026/03/27/openai-responses-streaming-backpressure-chunk-reassembly-timeout-budget/</guid>
      <description>&lt;p&gt;Most streaming failures are not about “can it stream”, but “does it stay stable under load”: broken chunks, stuck clients, timeout cascades, and retry storms.&lt;/p&gt;</description>
    </item>
    <item>
      <title>OpenAI Responses &#43; Go Stream Recovery: Delta Persistence, Resume Tokens, and Duplicate Chunk Dedup</title>
      <link>https://www.mfun.ink/en/2026/03/23/openai-responses-go-stream-resume-delta-dedup/</link>
      <pubDate>Mon, 23 Mar 2026 01:13:09 +0000</pubDate>
      <guid>https://www.mfun.ink/en/2026/03/23/openai-responses-go-stream-resume-delta-dedup/</guid>
      <description>&lt;p&gt;In production, the painful part is not “streaming is slow.” It’s “streaming breaks and then duplicates output after reconnect.”&lt;br&gt;
This guide gives you a practical recovery loop: &lt;strong&gt;delta persistence + resume token + idempotent dedup&lt;/strong&gt;, so reconnection does not replay garbage.&lt;/p&gt;</description>
    </item>
    <item>
      <title>OpenAI Responses in Go Multi-Tenant Quota Governance: Token Buckets, Budget Circuit Breakers, and Cost Attribution</title>
      <link>https://www.mfun.ink/en/2026/03/20/openai-responses-go-multitenant-quota-governance/</link>
      <pubDate>Fri, 20 Mar 2026 01:08:00 +0000</pubDate>
      <guid>https://www.mfun.ink/en/2026/03/20/openai-responses-go-multitenant-quota-governance/</guid>
      <description>&lt;p&gt;Most multi-tenant AI platforms fail for two boring reasons: one tenant saturates shared capacity, and finance discovers the burn too late.&lt;/p&gt;
&lt;p&gt;This guide gives you a practical Go blueprint: token-bucket throttling, budget circuit breakers, and request-level cost attribution.&lt;/p&gt;</description>
    </item>
    <item>
      <title>Go &#43; OpenAI Responses Agent Memory Layering: Short-Term Context, Long-Term Index, and Cost Caps</title>
      <link>https://www.mfun.ink/en/2026/03/18/go-openai-responses-agent-memory-layering/</link>
      <pubDate>Wed, 18 Mar 2026 16:33:52 +0000</pubDate>
      <guid>https://www.mfun.ink/en/2026/03/18/go-openai-responses-agent-memory-layering/</guid>
      <description>&lt;p&gt;In production Go agents, the first thing that breaks is usually not model quality. It is memory management: context grows, bills spike, and answers drift.&lt;/p&gt;
&lt;p&gt;Use a 3-layer memory design:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;L1: short-term conversational window (seconds)&lt;/li&gt;
&lt;li&gt;L2: rolling summary (minutes)&lt;/li&gt;
&lt;li&gt;L3: long-term retrieval memory (days)&lt;/li&gt;
&lt;/ul&gt;</description>
    </item>
    <item>
      <title>OpenAI Responses &#43; GitHub Actions PR Risk Gate: Automated Evals, Tiered Blocking, and One-Click Rollback</title>
      <link>https://www.mfun.ink/en/2026/03/16/openai-responses-github-actions-pr-risk-gate/</link>
      <pubDate>Mon, 16 Mar 2026 01:08:00 +0000</pubDate>
      <guid>https://www.mfun.ink/en/2026/03/16/openai-responses-github-actions-pr-risk-gate/</guid>
      <description>&lt;p&gt;You don&amp;rsquo;t need an AI reviewer that “sounds smart.” You need a gate that &lt;strong&gt;stops risky PRs before they hit main&lt;/strong&gt;.&lt;/p&gt;
&lt;p&gt;This post shows a production-ready minimum setup: OpenAI Responses generates structured risk output, GitHub Actions enforces tiered policies, and critical failures can trigger a one-click rollback.&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
