GEOTAB MCP CONNECTOR

You already run Geotab. Now ask it questions instead of building reports.

This is what it looks like when an AI assistant connects to your live MyGeotab data through Geotab's official MCP server — read fleet state, get diagnoses, and trigger real actions, all in plain English.

Plain-English fleet questions. Weekly reviews, fault triage, "who's closest and free right now" — answered from live data, not a static export.
Start with Claude, or bring your own assistant. MCP (the Model Context Protocol) is an open standard for connecting assistants to tools — Claude, Microsoft Copilot, ChatGPT, and other MCP-capable clients can all speak to the same Geotab server.
Try this version risk-free. Everything below runs against an anonymized Geotab demo fleet — no real account, no setup, no risk to production data.

Already comfortable with MCP? Jump straight in below.

Settings

Tune how the simulator behaves. More controls can live here as the demo grows.

Playback

Choose how fast the pre-recorded MCP transcript plays back.

Speed: xfast

Fleet progress game experimental

Off by default, so the simulator stays a simulator. Turn it on and the scenarios you explore grow a simulated fleet toward the full 50 vehicles — checkmarks and points included. Stored only in this browser.

off

About this site

Built by Felipe Hoffa — LinkedIn.

Learn more about the real connector on the Geotab MCP Connector landing page.

New to MCP? The Model Context Protocol is the open standard that lets AI assistants call external tools like MyGeotab — see modelcontextprotocol.io.

Found a bug, or want to see another scenario? Open a GitHub issue.

Connect your real Geotab account

Three things to do before an AI assistant touches your real fleet data.

Get access to a MyGeotab database. The fastest way to start is a free demo database — register at my.geotab.com/registration.html to practice on anonymized data first (same shape as this simulator, no real drivers or vehicles). For production access, ask your Geotab account rep or fleet admin.
Connect the Geotab MCP connector to your assistant of choice. Geotab publishes an official MCP (Model Context Protocol) server. Claude is the clearest starting point today; Microsoft Copilot, ChatGPT, and other MCP-capable clients can connect to the same server using the same URL below; only the menu names differ.
Start with Geotab's official walkthrough: Getting started with MyGeotab MCP. The quick version is below.

Example client setup:
1. Go to claude.ai → Settings → Connectors, then click Add custom connector.
2. Fill in a name (e.g. "Geotab") and the remote MCP server URL: https://mcp.geotab.com/mygeotab. Leave the OAuth fields blank — Geotab's server handles that for you.
4. Click Add — Claude redirects you to a MyGeotab sign-in screen. Log in with your MyGeotab username and password and approve the connection.
5. Back in the client, the connector now shows as Connected. Turn it on for a chat from the tools/connectors picker (the connector icon below the message box), then ask it anything you tried in this simulator.
Microsoft Copilot, ChatGPT, and other MCP-capable clients follow the same pattern — add a custom/remote MCP connector, point it at https://mcp.geotab.com/mygeotab, and sign in with your MyGeotab credentials when prompted.
Know what you're exposing.
Before connecting a production database: the connector inherits your MyGeotab permissions, so the assistant can see anything you can see — including personal data (driver names, employee numbers, emails, precise location history).
- Review your AI provider's DPA and your GDPR / data-minimization obligations.
- Instruct the assistant to work at the vehicle/asset level — ask it to avoid driver names, emails, and phone numbers in its answers.
- Demo databases (like the one above) are anonymized — "Demo - 01", no real driver data — so they're safe to explore freely.
Not legal advice — check with your own counsel/compliance team before connecting production data.