From 8845d22ce9ad4e9f6199631c6fd02c97444146d1 Mon Sep 17 00:00:00 2001
From: "russell@unturf.com" <russell@unturf.com>
Date: Tue, 24 Feb 2026 18:07:21 -0500
Subject: [PATCH] document Reverse Retrieval Augmented Generation as the core
 technique in uncloseai.js

---
 public/uncloseai-js-styleguide.html |  2 +-
 public/uncloseai-js.html            | 12 ++++++++----
 2 files changed, 9 insertions(+), 5 deletions(-)

diff --git a/public/uncloseai-js-styleguide.html b/public/uncloseai-js-styleguide.html
index 8eca4cd..7bfe83e 100644
--- a/public/uncloseai-js-styleguide.html
+++ b/public/uncloseai-js-styleguide.html
@@ -797,7 +797,7 @@ curl -X POST https://speech.ai.unturf.com/v1/audio/speech \
         <h3 id="embed-code">Embed Code</h3>
         <pre><code class="language-html">&lt;!-- Add to any HTML page --&gt;
 &lt;script src="https://uncloseai.com/uncloseai.js" type="module"&gt;&lt;/script&gt;</code></pre>
-        <p>That single line loads the entire modal system: vault encryption, chat, TTS, translation, page reading, file upload, and settings. Zero configuration required. See the <a href="/uncloseai-js.html">full documentation</a> for customization options.</p>
+        <p>That single line loads the entire modal system: vault encryption, chat with Reverse RAG (live page content injected into every conversation), TTS, translation, file upload, and settings. Zero configuration required. See the <a href="/uncloseai-js.html">full documentation</a> for customization options.</p>
 
         <footer>
         <small>© uncloseai. 2026</small>
diff --git a/public/uncloseai-js.html b/public/uncloseai-js.html
index 94bc7ab..a05c8d6 100644
--- a/public/uncloseai-js.html
+++ b/public/uncloseai-js.html
@@ -103,6 +103,8 @@
         <h2 id="intro">uncloseai.js</h2>
         <p><strong>uncloseai.js</strong> is a single-file JavaScript library that adds a floating chat button to any website. One script tag gives your page a full machine learning interface: streaming chat, text-to-speech, page reading, translation, file upload, encrypted settings, and 19 languages. No server, no backend, no API keys.</p>
 
+        <p>The core technique is <strong>Reverse Retrieval Augmented Generation (Reverse RAG)</strong>: instead of a server fetching documents to stuff into a prompt, the client extracts live page content and injects it directly into the conversation context. The model always has the freshest data from the page you're viewing, which lets small models punch way above their weight class. No vector database, no embeddings, no indexing pipeline. Just real-time content from the DOM, straight into the prompt.</p>
+
         <p>Works with static sites, CDNs, GitHub Pages, and any hosting that serves HTML. Every feature runs client-side against free community inference endpoints.</p>
 
         <!-- ============================================ -->
@@ -126,7 +128,7 @@
             <li>Copy-to-clipboard button on every code block</li>
             <li>Per-page conversation history, encrypted and persisted</li>
             <li>Multi-model support with dynamic model discovery from endpoints</li>
-            <li>Context-aware page reading: the model sees your page content automatically</li>
+            <li>Reverse RAG: live page content extracted and injected into every conversation for context-aware responses</li>
         </ul>
 
         <h3 id="tts">Text-to-Speech</h3>
@@ -147,9 +149,11 @@
             <li>Opens translated content in a new browser tab</li>
         </ul>
 
-        <h3 id="page-reading">Page Reading</h3>
+        <h3 id="page-reading">Page Reading (Reverse RAG)</h3>
         <ul>
-            <li>Extracts main content from any webpage intelligently</li>
+            <li>Extracts main content from any webpage intelligently, straight from the live DOM</li>
+            <li>Injects page content directly into the conversation context (Reverse RAG): no server-side retrieval, no vector store, no embeddings</li>
+            <li>Small models get the full picture of whatever page you're on, producing answers that rival much larger models</li>
             <li>Synthesizes a short summary with the machine learning model</li>
             <li>Reads the summary aloud with selected voice</li>
             <li>Play, pause, and download the audio</li>
@@ -508,7 +512,7 @@ window.uncloseai.UncloseVault.lock();</code></pre>
                             <li><a href="#chat">Chat</a></li>
                             <li><a href="#tts">TTS</a></li>
                             <li><a href="#translation">Translation</a></li>
-                            <li><a href="#page-reading">Page Reading</a></li>
+                            <li><a href="#page-reading">Page Reading (Reverse RAG)</a></li>
                             <li><a href="#vault">Vault</a></li>
                             <li><a href="#file-upload">File Upload</a></li>
                             <li><a href="#code-execution">Code Execution</a></li>