Model Context Protocol

Give your AI agent shipment tracking

Shipyu has an official MCP server. Connect it once and Claude Code, Cursor, Codex, or any MCP client can create trackers, read live shipment status, manage webhooks, and check your usage. Test keys are free — an agent can verify the whole integration without spending money.

Claude Code — one command

No API key to paste. Your browser opens to sign in, where you choose the organization and whether the agent gets test or live access. Sign up free — test mode costs nothing.

claude mcp add shipyu --scope user --transport http https://shipyu.com/api/mcp

Then try it:

Track 9400111899560000000000 with USPS and tell me where it is.

Cursor and other MCP clients

Add this to your mcp.json. The server speaks MCP Streamable HTTP and signs you in through the browser. A client that cannot open a browser can send a Shipyu API key as a Bearer token instead.

{
  "mcpServers": {
    "Shipyu": {
      "type": "http",
      "url": "https://shipyu.com/api/mcp"
    }
  }
}

16 tools, the full tracking workflow

get_docs	Read a Shipyu documentation topic
list_carriers	List the carriers Shipyu can track, with the exact id to pass as `carrier` on create_tracker
create_tracker	Start tracking a shipment
get_tracker	Retrieve a tracker by its Shipyu tracker id (trk_...)
list_trackers	List trackers YOUR account created, newest first
delete_tracker	Stop tracking a shipment and delete the tracker by id
list_webhooks	List all webhook endpoints registered for this API client, with status and failure counts
create_webhook	Register a webhook endpoint
update_webhook	Change a registered webhook's URL, description, subscribed event types, or status
delete_webhook	Permanently delete a webhook endpoint and its delivery history
test_webhook	Send a signed sample tracker.updated event to a registered webhook endpoint and report the HTTP outcome
list_webhook_deliveries	Show what Shipyu tried to deliver to one webhook endpoint and what that endpoint returned on each attempt, including HTTP status, response body, and the next scheduled retry
list_requests	List this account's recent API requests with status, latency, and error code
get_request	Read one recorded API request by its request id, including the exact request and response bodies Shipyu stored
diagnose_tracker	Explain why a shipment is not moving
get_usage	Get this connection's identity and usage: the account name, whether it is in test or live mode, the key it resolved to, rate limits, current month totals, monthly history, and the most recent requests

Common questions

Does Shipyu have an official MCP server?+

Yes. Shipyu runs an official Model Context Protocol server at https://shipyu.com/api/mcp. It speaks MCP Streamable HTTP and gives an AI agent 16 tools for shipment tracking, including creating trackers, reading live carrier status, managing signed webhooks, reading the API documentation, and inspecting its own failed requests.

How do I connect Shipyu to Claude Code?+

Run: claude mcp add shipyu --scope user --transport http https://shipyu.com/api/mcp. Claude Code then opens your browser to sign in to Shipyu, where you choose the organization and whether the agent gets test or live access. There is no API key to copy or paste.

Do I need an API key to use the Shipyu MCP server?+

No. The server supports OAuth 2.1, so your MCP client opens a browser sign-in and is issued its own credential. Clients that cannot open a browser can still authenticate by sending a Shipyu API key as a Bearer token.

Is the Shipyu MCP server free to use?+

The server and its tools are free. The only billable action is creating a tracker with a live API key, which starts at $0.020 per tracker and falls at volume. Test mode is free, so an agent can build and verify an entire integration without spending anything.

Which MCP clients work with Shipyu?+

Any client that speaks MCP Streamable HTTP. Shipyu is used with Claude Code and Cursor, and the same URL works in other MCP clients. The server supports 5 protocol revisions, so a client on an older revision keeps working without changes.

What can an AI agent do with Shipyu shipment tracking?+

It can create a tracker from a carrier tracking number, read the normalized status and the full scan timeline, register and test signed webhooks, look up which carriers are supported, read the API documentation, and diagnose why a specific shipment has stopped moving.

Can an AI agent verify Shipyu webhook signatures?+

Yes. The server ships a webhook-verification documentation topic giving the exact header format, the signed string, the tolerance window, and working Node and Python examples. The test_webhook tool then sends a real signed delivery, so the agent can confirm its verification code accepts a genuine Shipyu signature.

How does an agent learn the Shipyu API without leaving the chat?+

The get_docs tool serves 12 documentation topics generated from the live OpenAPI specification and the running configuration, so they cannot drift from the API. Each topic is also published as an MCP resource, which lets you attach it to a message directly in clients that support resource mentions.

Start with a free test key

Sign up, run the command above, sign in when your browser opens. Test mode returns realistic sample data and costs nothing. Live tracking starts at $0.020 per tracker and falls with volume.

Get a free test key API reference