diff --git a/.github/workflows/pages.yml b/.github/workflows/pages.yml
new file mode 100644
index 0000000..6b81519
--- /dev/null
+++ b/.github/workflows/pages.yml
@@ -0,0 +1,23 @@
+name: pages
+on:
+  push: {branches: [main]}
+  workflow_dispatch:
+permissions: {contents: read, pages: write, id-token: write}
+concurrency: {group: pages, cancel-in-progress: true}
+jobs:
+  build:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - run: mkdir -p _site && cp -r site/* _site/ && cp -r docs/screenshots _site/ && touch _site/.nojekyll
+      - uses: actions/upload-pages-artifact@v3
+        with: {path: _site}
+  deploy:
+    needs: build
+    runs-on: ubuntu-latest
+    environment:
+      name: github-pages
+      url: ${{ steps.d.outputs.page_url }}
+    steps:
+      - id: d
+        uses: actions/deploy-pages@v4
diff --git a/README.md b/README.md
index 6af8572..7dea6ca 100644
--- a/README.md
+++ b/README.md
@@ -1,5 +1,7 @@
 # AgentOps MCP
 
+**Live demo (static, GitHub Pages):** https://naveecode.github.io/agentops-mcp/
+
 [![ci](https://github.com/naveecode/agentops-mcp/actions/workflows/ci.yml/badge.svg)](https://github.com/naveecode/agentops-mcp/actions)
 
 A deployable AI agent service. A LangGraph ReAct agent answers questions over your documents and calls tools. The same tools are exposed as an **MCP server**, so Cursor and Claude Desktop can use them. Documents are chunked, embedded, and stored in MongoDB. Ingest runs inline, or through a RabbitMQ worker that retries and then dead-letters.
diff --git a/site/index.html b/site/index.html
new file mode 100644
index 0000000..e2490fe
--- /dev/null
+++ b/site/index.html
@@ -0,0 +1,17 @@
+<!doctype html><html><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1">
+<title>AgentOps MCP demo</title><style>body{font-family:system-ui;max-width:960px;margin:2rem auto;padding:0 1rem;color:#111}img{max-width:100%;border:1px solid #ddd;border-radius:8px}.row{display:flex;gap:1rem;align-items:flex-start;flex-wrap:wrap}.row img{flex:1;min-width:240px}code{background:#f3f3f3;padding:0 .3rem}</style>
+<script type="module">import m from "https://cdn.jsdelivr.net/npm/mermaid@11/dist/mermaid.esm.min.mjs";m.initialize({startOnLoad:true});</script></head><body>
+<h1>AgentOps MCP</h1>
+<p>A tool-calling LangGraph agent served by FastAPI with a Redis cache. The same tools are exposed as an MCP server for Cursor and Claude Desktop. <a href="https://github.com/naveecode/agentops-mcp">Source on GitHub</a>.</p>
+<p><em>This is a static demo page. The live backend is not hosted; run it locally with <code>docker compose up</code>.</em></p>
+<h2>Chat UI</h2><div class="row"><img src="screenshots/chat-desktop.png" alt="Chat UI, desktop"><img src="screenshots/chat-mobile.png" alt="Chat UI, mobile" style="max-width:280px"></div>
+<h2>Architecture</h2><pre class="mermaid">flowchart LR
+  UI[Next.js chat UI] --> API[FastAPI /ask]
+  API --> C[(Redis cache)]
+  API --> AG[LangGraph ReAct agent]
+  AG --> T[Tools: search_docs, calculator]
+  MCP[MCP server stdio] --> T
+  API --> K[Docker / kind K8s]</pre>
+<h2>Quickstart</h2><pre>git clone https://github.com/naveecode/agentops-mcp && cd agentops-mcp
+docker compose up        # API on :8000, Redis
+python -m app.mcp_server # MCP for Cursor / Claude</pre></body></html>