<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Sse on Step Into Dev</title><link>https://stepinto.dev/tags/sse/</link><description>Recent content in Sse on Step Into Dev</description><generator>Hugo</generator><language>en-gb</language><lastBuildDate>Mon, 23 Mar 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://stepinto.dev/tags/sse/index.xml" rel="self" type="application/rss+xml"/><item><title>Streaming AI Responses from Server to Screen</title><link>https://stepinto.dev/posts/streaming-structured-responses/</link><pubDate>Mon, 23 Mar 2026 00:00:00 +0000</pubDate><guid>https://stepinto.dev/posts/streaming-structured-responses/</guid><description>&lt;p&gt;I wrote previously about &lt;a href="https://stepinto.dev/posts/streaming-json-partial-parser/"&gt;a partial JSON parser&lt;/a&gt; for extracting text from incomplete LLM responses as they stream in. That post covered one specific piece: pulling the &lt;code&gt;response&lt;/code&gt; field out of a half-received JSON object so users see text immediately.&lt;/p&gt;
&lt;p&gt;But that parser lives inside a larger system. There&amp;rsquo;s a server emitting tokens over SSE, a chunk protocol that handles tool execution mid-response, a transport layer that works on React Native (barely), and a rendering pipeline that switches from plain text to rich content when the stream completes. The parser was the fun bit to write about. The pipeline around it is where the real engineering decisions live.&lt;/p&gt;</description></item></channel></rss>