Carriyo MCP.
Carriyo exposes a Model context protocol server so AI agents and LLM-based applications can read from — and act on — your Carriyo data using natural language.
MCP is not a different product surface; it's a different interface to the same product. Tools are designed to be invoked by an LLM — schemas are agent-friendly, descriptions are richer, and responses are structured to be summarisable rather than just deserialisable.
The agent acts on your behalf. When you authorize an agent (see Getting started), it inherits your Dashboard user's permissions and tenant / merchant scope. Whatever you can do, the agent can do.
What you can do with it
The Carriyo MCP server exposes 39 tools across 5 families.
A typical agent doesn't call one tool — it chains several. For canonical chains, see Prompts & patterns.
When to use MCP vs the REST API
If a human or pre-built integration is calling Carriyo, use the REST API. If an LLM is composing the call from natural-language intent — agentic workflows, copilots, internal AI assistants — use MCP.
| Scenario | Use |
|---|---|
| E-commerce platform calling Carriyo | REST API |
| WMS / OMS integration | REST API |
| ChatOps bot ("show me yesterday's failed deliveries") | MCP |
| Internal copilot for ops team | MCP |
| Agent that reads emails and creates shipments | MCP |
| Public chat that quotes and books for end customers | MCP |
The two interfaces are interchangeable in capability — anything you can do with one, you can do with the other (with the exception of a few REST-only admin operations). Pick the interface that matches how the call is composed.