Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@MCP E-Commerce Agent Show me the latest pending orders"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A proof of concept that uses the Model Context Protocol (MCP) to give an e-commerce company's AI agents the business context they need to run operations — live orders, inventory, logistics, returns, claims, and payments — safely and at scale.
📖 Full documentation: https://yassineteimi.github.io/mcp-ecommerce-agent/
A general LLM doesn't know that order #88213 is late because its driver called in sick. MCP connects AI agents to real business systems through one open standard, turning a brittle N×M web of integrations into a reusable N+M model: wrap each system once as an MCP server, and any agent can use it.
The PoC follows a deliberate maturity path:
Internal first — a supply-chain manager gets an assistant with governed, read-mostly access to every fulfilment system; the agent proposes actions, a human approves them.
Mature — enable trusted write actions (reorder, reroute, approve return) behind guardrails, audit, and evaluation.
Customer-facing — expose a curated, per-tenant-isolated subset (order tracking, returns) into the customer experience.
Related MCP server: merch-connector
Servers, one per domain — Orders, Inventory, Logistics, Returns, Claims, Payments — each fronting a system of record with isolated credentials.
Three MCP primitives — Resources (read-only context), Tools (human-approved actions), Prompts (reusable operating procedures).
Built with the Python MCP SDK (FastMCP) — a complete server is ~15 lines; the same code runs locally over stdio or remotely over Streamable HTTP.
Secure by design — OAuth 2.1 authorization, role-based scoped access, per-server least privilege, transport hardening, and per-tenant isolation for the customer phase.
Page | Contents |
What MCP is and why it matters for AI-agent integration | |
Servers, clients, and the Tools/Resources/Prompts flow (with diagrams) | |
How the servers and clients are built | |
Authorization, scoped access, transport hardening |
Docs are built with MkDocs Material and deployed to GitHub Pages by .github/workflows/ci.yml on every push to main.
A design-led proof of concept: the architecture, code patterns, and security model are production-shaped and standards-accurate, illustrated against a representative e-commerce domain. Learn more about MCP at modelcontextprotocol.io.
This server cannot be deployed
Hosted MCP for e-commerce: live product catalog, stock, and pricing for AI agents.
The Mercado Pago MCP Server implements the Model Context Protocol to provide AI agents and LLMs with access to Mercado Pago's APIs and tools within compatible development environments. It acts as an intermediary that translates Mercado Pago resources into executable functions (tools) that AI applications can invoke to perform actions and automate flows. The server simplifies integration, enables using documentation to implement or improve code, and optimizes operations through natural language interactions without manual implementations.
MCP server unifying ERPs, CRMs, APIs and knowledge base for Claude, ChatGPT and Gemini.
Multi-tenant MCP gateway for AI commerce. One connection, every store.
