How to Track Containers with ChatGPT and Claude (MCP Server Guide)

What is MCP?

Model Context Protocol (MCP) is an open standard that lets AI assistants call external tools. Think of it as a USB-C port for AI: a single protocol that any LLM can use to query any data source. Anthropic released MCP in late 2024; by mid-2026 it's supported by Claude, Cursor, ChatGPT, and most agentic frameworks.

Why container tracking via MCP?

If you're a logistics manager using Claude to draft emails, summarize meeting notes, or plan shipments, you'd love to ask: 'When does MSCU1234567 arrive in Long Beach?' Without MCP, the AI can't help. With MCP, the AI calls MariMetric's track_container tool and answers in plain English with the actual ETA.

Installation

1. 1Install Node.js 18+ (skip if you have it)
2. 2Get a MariMetric API key (free at marimetric.com/dashboard/api-keys)
3. 3Edit ~/.config/claude-desktop/claude_desktop_config.json (or equivalent for your MCP client):
4. 4Add the MariMetric MCP server entry (see code snippet below)
5. 5Restart your MCP client. You should see 'mari_metric' as an available server.

Example config

{
  "mcpServers": {
    "mari_metric": {
      "command": "npx",
      "args": ["-y", "@marimetric/mcp-server"],
      "env": {
        "MARIMETRIC_API_KEY": "your-api-key"
      }
    }
  }
}

Example queries

• Track MSCU1234567
• Show me all containers in the trans-Pacific fleet that are delayed
• What's the average transit time Shanghai → Rotterdam this month?
• Get vessel info for IMO 9776418
• Validate this container number: MSCU1234567

Roadmap

V2.5 of MariMetric will launch the public MCP server. For early access, contact us at [email protected]. Pricing for MCP calls follows our regular tiers: 100 calls/month free, then call packs or subscription plans.