diff --git a/README.md b/README.md
index d2f2998..31b785f 100644
--- a/README.md
+++ b/README.md
@@ -7,6 +7,8 @@
 [![SARIF 2.1.0](https://img.shields.io/badge/output-SARIF%202.1.0-orange)](#ci--enterprise-integration)
 [![License: MIT](https://img.shields.io/badge/license-MIT-lightgrey)](LICENSE)
 
+A senior pentester's "here's what to test and how" handoff — auto-generated from your repo, for your AI agent to execute.
+
 <!-- docguard:quality negation-load off — "no LLM / no server / no running app / not a SaaS / never touches prod" is this tool's core positioning; defining it by contrast with the scanners-and-SaaS it deliberately is NOT is intentional, not a phrasing defect. -->
 
 > Local-first security recon that **briefs your AI coding agent**. It does the deterministic
@@ -15,6 +17,8 @@
 > Gemini, Cursor) a marching-orders briefing. **Code in, artifacts out. No LLM in the tool, no
 > server, no running app required.**
 
+[![websec-validator demo](assets/demo.gif)](assets/demo.gif)
+
 It is *not* an autonomous scanner and *not* a SaaS. It's the missing front-half: the thing that
 turns a repo into a precise, fact-grounded security brief an AI agent (with a human in the loop)
 can act on — an auto-filled, repo-aware version of a senior pentester's "here's what to test and
diff --git a/assets/demo.gif b/assets/demo.gif
new file mode 100644
index 0000000..89237f2
Binary files /dev/null and b/assets/demo.gif differ
diff --git a/assets/demo.tape b/assets/demo.tape
new file mode 100644
index 0000000..6619e83
--- /dev/null
+++ b/assets/demo.tape
@@ -0,0 +1,37 @@
+# demo.tape — regenerates assets/demo.gif with charmbracelet/vhs.
+#
+# Render:
+#   1. Install websec so `websec` is on PATH (e.g. `pipx install websec-validator`).
+#   2. Clone the demo target:  git clone --depth 1 https://github.com/erev0s/VAmPI /tmp/VAmPI
+#   3. From the repo root:      vhs assets/demo.tape
+#
+# Requires: vhs + its ttyd/ffmpeg deps (`brew install vhs`).
+# VAmPI is a deliberately-vulnerable API from our proof-harness corpus, so
+# scanning it is expected use. `websec run` is read-only recon (no --scan here).
+
+Output assets/demo.gif
+
+Set Shell "bash"
+Set FontSize 20
+Set Width 1200
+Set Height 640
+Set Theme "Catppuccin Mocha"
+Set TypingSpeed 40ms
+Set Framerate 24
+Set Padding 20
+
+# Clean, minimal prompt — hidden from the recording.
+Hide
+Type "export PS1='$ '" Enter
+Type "clear" Enter
+Show
+
+# 1) Point it at a repo — deterministic recon maps the whole attack surface in ~1s.
+Type "websec run /tmp/VAmPI"
+Enter
+Sleep 4s
+
+# 2) The product: the marching-orders briefing your AI agent executes.
+Type "head -48 websec-out/latest/AGENT-BRIEFING.md"
+Enter
+Sleep 6s