Run the Senserva MCP in demo mode

Drive Siemserva from Claude against a built-in, fictional Microsoft 365, Intune, Defender, and Entra ID tenant, in plain language. No registration, no key, and no access to your own tenant. One command builds the demo database and wires it into Claude for you.

1

Download Siemserva by Senserva, free

Grab the signed build for your OS. One small binary, no installer, no agents.

Windows (x64) macOS (Apple Silicon)
2

Build the demo and wire up Claude, one command

Run demo claude. Pick a tenant size; Siemserva builds the demo database and automatically writes the MCP config for both Claude Desktop and Claude Code. No config files to edit by hand.

Already connected the MCP? You do not even need the command line, just ask Claude to “set up the Siemserva demo” and it builds and loads the database for you.

> siemserva-win-x64.exe demo claude
$ ./siemserva-osx-arm64 demo claude
3

Install the free Claude Desktop app

If you do not have it yet, install Claude Desktop (or Claude Code). The config from step 2 is already in place, so whichever you install picks up Siemserva on startup. You bring your own model; there is no AI markup.

4

Open Claude and ask in plain language

Open or restart Claude Desktop (or run claude), then ask. Claude calls the Senserva MCP against the demo data and answers from your real scan, with the evidence behind it.

> which admins have no MFA?
> summarize the top 5 risks in my Siemserva scan
> what CVEs affect my Windows fleet?

Want a challenge instead of a chat? Just say “want to play a game” and play Game Mode against the same demo data.

Prefer the dashboard? Run siemserva-win-x64.exe demo dashboard for the live TUI instead. Ready to scan your own tenant? Register free. For the full setup, see the Quick Start.

What is actually happening

Demo mode generates a complete, fictional tenant and runs the same 650+ checks the product runs in production, then exposes the results to Claude through the Senserva MCP. Claude is not reading the open web or guessing: every answer is grounded in the demo scan's findings and the one connected model of configuration, patching, and logs, and validated through Senserva Trustworthy AI. Want a graded challenge against Claude on the same data? Play You v. Claude.

Claude and the MCP