Connect Cargofy To ChatGPT, Claude, Gemini And Other LLMs Via MCP
Cargofy now has an MCP server, so you can connect it to the AI assistant your team already uses: ChatGPT, Claude, Gemini, Perplexity or Manus. Ask about your shipments in plain words and get answers from live account data. The assistant can also act in Cargofy for you, with your permissions and nothing more. Setup is copy, paste, confirm. No code.
Most logistics teams already use ChatGPT or Claude for something. The workflow usually looks like this: export a CSV, clean it up, paste it into the chat, ask the question, then check the answer against the system because the export is already a day old. With the connection in place, the LLM reads your logistics data in Cargofy directly. It works alongside the Architect, the AI already built into Cargofy, and brings in the context the Architect doesn't have.
What you can ask your AI assistant about shipments
The assistant sees what you see in Cargofy. Type "Show me the last three shipments" and it pulls them from your account: routes, dates, status.
Bigger asks work too. In our internal test we asked Claude for "a chart of all shipments by type, date and route, with a mini map". It read every shipment in the account and returned an interactive report. A report like this takes longer than a quick lookup.
The assistant can also do in Cargofy what you can do by hand: create or update records, check a shipment, pull documents. Ask in words and it finds the right place in the platform.
The real gain comes when one assistant holds several of your tools. If it is also connected to your email or Google Sheets, one request covers both: find this week's shipments in Cargofy, then draft the customer updates or fill the weekly report.
The Architect and your own LLM work together
Cargofy already has an AI built in: the Architect. Open it inside the platform and ask in plain words. It sets up workflows, Digital Workers, Superpowers and knowledge bases from a conversation, hands business requests to the right Digital Worker and shows the result in the same chat. Everything it creates stays switched off until a person approves it.
For work inside Cargofy, start with the Architect. It knows the platform and your account setup, and it needs no connection.
Connect your own LLM when the task needs context that lives outside Cargofy: your email, contracts, Google Sheets, notes from a customer call. Or when your team already spends the day in ChatGPT or Claude and wants Cargofy data there too.
One task can use both. In Claude, you compare this month's shipments in Cargofy with the contract rates from a carrier's email and find the lanes where carriers bill above contract. Then you open the Architect and ask it to set up a workflow that flags those lanes on every new tender.
Which LLMs and AI agents connect to Cargofy
Cargofy connects to Claude, ChatGPT, Gemini, Perplexity and Manus. It also connects to the agent platforms Paperclip, OpenClaw and Hermes, if your team runs its own AI agents for logistics.
Cargofy works as an MCP server. MCP (Model Context Protocol) is the open standard that ChatGPT, Claude and other AI assistants use to connect to outside tools and data. You don't need to know how it works to use it.
How to connect Cargofy to ChatGPT, Claude or Gemini
- In Cargofy, open MCP & API in the System block of the main menu and copy the connection address.
- In ChatGPT, Claude or another assistant, add a custom MCP connector and paste the address.
- Click Connect. Cargofy opens a confirmation screen. Approve access, and the assistant is connected.
The MCP & API page has step-by-step instructions for each assistant, with no commands to type. The same steps are in the Cargofy Documentation.
Security: the AI assistant works with your permissions
The assistant works in Cargofy on behalf of one person: inside that person's company, with that person's role and permissions. If your role can't publish a workflow, your assistant can't either.
Every action it takes shows up in the history as yours, marked with the assistant that made it. When a colleague asks who changed a shipment, the log answers: you, through Claude, and when.
The MCP & API page lists every connected assistant with the date it was connected. To disconnect one, click Revoke. You can do it any time.
Cargofy API and webhooks for technical teams
The API key, webhooks and the webhook log now live on the second tab of the same MCP & API page. Old links from account settings redirect there.
FAQ
What is an MCP server?
An MCP server lets an LLM read data from a platform and take actions in it. Cargofy's MCP server gives your AI assistant access to your shipments, tenders and documents, within your own permissions.
Can I connect ChatGPT to my logistics platform without a developer?
Yes. Copy one address from Cargofy, paste it into ChatGPT, Claude or Gemini and confirm access on the Cargofy screen. No code and no API key.
Which AI assistants work with Cargofy?
Claude, ChatGPT, Gemini, Perplexity and Manus, plus the AI agent platforms Paperclip, OpenClaw and Hermes.
Does connecting an LLM replace the Architect?
No. The Architect stays the AI inside Cargofy that sets up workflows and Digital Workers. Your own LLM adds context from outside Cargofy, such as email, documents and spreadsheets, and lets you work with Cargofy data where your team already works.
Can the AI assistant see data outside my company?
No. It works inside your company account, with your role and permissions. It sees and does only what you can see and do by hand.
How do I disconnect an AI assistant from Cargofy?
Open MCP & API in the System block of the main menu and click Revoke next to the assistant.
Try it
Already on Cargofy? Open MCP & API and connect the assistant you use every day. Not on Cargofy yet? Book a demo.