diff --git a/docs/marketing/devrel/demos/failed-workspace-ec2-console-demo.md b/docs/marketing/devrel/demos/failed-workspace-ec2-console-demo.md
new file mode 100644
index 0000000..cd96a31
--- /dev/null
+++ b/docs/marketing/devrel/demos/failed-workspace-ec2-console-demo.md
@@ -0,0 +1,66 @@
+# Failed Workspace EC2 Console Output — Demo Script
+**Source:** PR #1178 | **Acceptance:** Working demo + repo link + 1-min screencast
+
+---
+
+## What This Demo Shows
+
+1. Trigger a workspace to fail (bootstrap error, bad config, missing env vars)
+2. Open the failed workspace in Canvas
+3. Open the error panel — EC2 console output is visible without leaving Canvas
+4. Compare: what you used to have to do (open AWS Console, find instance, pull logs)
+
+**Time:** ~90 seconds | **Prereqs:** Self-hosted Molecule AI with EC2 provisioning, or a deployment that has triggered a failed workspace recently
+
+---
+
+## Demo Script
+
+### Step 1: Navigate to a Failed Workspace
+
+Open Canvas. Find a workspace in `failed` state — it will have a red status indicator in the workspace list.
+
+Click into the workspace.
+
+> **On screen:** Red "Failed" badge on the workspace card. Status panel shows last state transition.
+
+### Step 2: Open the Error Panel
+
+In the workspace detail view, locate the error panel (or "Last Error" section).
+
+> **On screen:** The error panel displays `last_sample_error` — a structured summary of what the provisioner observed.
+
+### Step 3: Reveal EC2 Console Output
+
+Click to expand the EC2 console output section within the error panel.
+
+> **On screen:** Full EC2 console log from the instance at the time of failure. Scrollable, timestamped entries.
+
+> **Narrator:** "You can see exactly what the EC2 instance recorded at the moment it failed — boot errors, service exits, OOM kills, anything that went to the console. No AWS Console tab required."
+
+### Step 4: (Optional) Compare the Old Way
+
+If time allows, briefly show the alternative:
+
+1. Open AWS Console in a new tab
+2. Navigate to EC2 → Instances
+3. Find the instance ID from the workspace metadata
+4. Open Instance Settings → Get System Log
+
+> **Narrator:** "This is what debugging a failed workspace looked like before today's release. Two tools, two contexts, a lot of tab-switching."
+
+---
+
+## Screencast Notes
+
+- Start recording before you open Canvas
+- Highlight the red "Failed" badge as the entry point
+- Expand the EC2 console output section slowly — let viewers read the log entries
+- Mention the instance ID or timestamp to show it's real output from the actual instance
+- Keep the AWS Console comparison brief — it's contrast, not the main event
+
+**Do not include real AWS credentials or instance IDs in the recording.**
+
+---
+
+*Attach to Molecule-AI/internal#7 for DevRel review.*