Connect Claude or ChatGPT to Orbit over MCP with your operator login: setup values, step-by-step connector configuration, and how permissions keep every action scoped to you.
Connect an AI assistant such as Claude or ChatGPT directly to Orbit. Once connected, your assistant can look up and update your Orbit data in plain language, always under your own login and permissions.
Orbit provides a connector endpoint based on the Model Context Protocol (MCP), an open standard supported by the major AI assistants. After a one-time setup, your assistant signs in with your normal Orbit operator account and can then work with your data: reading tours, creating carriers, checking shipments, and more.
Key highlights
Works with claude.ai (web, desktop, and mobile) and ChatGPT.
You sign in with your existing Orbit MissionControl login, including single sign-on.
Every action runs as you. Your Orbit roles and permissions apply to each request.
A workspace administrator can set the connector up once for the whole team.
Open claude.ai and go to Settings, then Connectors.
Select Add custom connector.
Enter the Server URL from the table above.
Open Advanced settings and enter the OAuth client ID, and in case future versions require it, the Authorization URL, and the Token URL.
Save, then select Connect. A sign-in window opens; log in with your Orbit operator account.
If your organisation uses Claude for Teams or Enterprise, an administrator can add the connector once in the workspace settings. Each member then only clicks Connect and signs in as themselves.
In ChatGPT, open Settings, then Connectors, and enable developer mode if asked.
Add a new connector with the same values from the table above.
Select Connect and sign in with your Orbit operator account.
If the sign-in fails with a redirect error, contact support@orbit.do. Some ChatGPT versions use a connector-specific return address that the Orbit team must approve once.
The assistant acts with exactly your permissions, never more. Anything you cannot do in Orbit MissionControl, the assistant cannot do either.
The full tool list is visible to every connected user, but each call is checked individually. A tool outside your permissions returns a clear "not permitted" message.
Assistants can change data, not just read it. Claude and ChatGPT ask for your confirmation before running an action; read those prompts before approving, exactly as you would review your own change.
Disconnecting the connector in your assistant's settings ends its access immediately.
Julia is an operator at Spaceport Shipping Co. and has connected Claude to Orbit. She types: "Which tours leave Paris tomorrow, and which of them still have no carrier?" Claude looks up tomorrow's Tour records and lists the unassigned ones. Julia follows up: "Create a carrier called Orion Freight Ltd with the contact details from this email." Claude shows her the create action for confirmation, she approves, and the new Carrier appears in Orbit MissionControl.
claude.ai in the browser, Claude Desktop, the Claude mobile apps, and ChatGPT. Developer tools that sign in from the command line, such as Claude Code, are not supported at this time.
Who can connect?
Any operator with an Orbit MissionControl login. Access for shippers, carriers, and drivers is not available through this connector.
Is this the same as the Orbit API?
No. The Orbit API uses API keys and is meant for system integrations; see the Orbit API Reference. The connector described here is for interactive AI assistants and uses your personal login. For AI coding agents that build against the API, see the Automating Orbit with AI Agents article.
Why do I see tools I cannot use?
The tool list is shared by all users, but every call is checked against your own permissions. Seeing a tool does not mean you can run it.
Can I limit what the assistant is allowed to do?
Yes, through your normal Orbit role. An operator with a read-only role gets a read-only assistant. Ask your administrator to adjust roles in Orbit MissionControl if needed.
