Skip to main content

Etalasse MCP Server

Etalasse provides a Model Context Protocol (MCP) server — an open standard that lets AI assistants and external applications interact with your shop data and operations through well-defined tools. The server runs over HTTP and supports OAuth authentication, so AI clients can connect to Etalasse like a first-class app.

Who this is for​

The Integrations section is aimed at developers and integrators:

  • teams connecting an AI client (such as Claude Code, Codex, or a ChatGPT connector) to Etalasse
  • teams building their own application or AI agent that calls Etalasse capabilities over MCP HTTP

If you are a merchant who simply wants to use AI inside the product, there is nothing to set up here — use the AI Workspace in the Dashboard.

What it can do​

Through the MCP server, an authenticated client can run Etalasse operations programmatically, including:

  • reading shop data — products, categories, orders, customers, and reports
  • managing the catalog — creating and updating products, variants, images, and add-ons
  • running operations — updating order status, managing bookings, stock, promotions, and withdrawals
  • serving storefront flows — customer/guest sessions, carts, checkout, and digital asset access

The MCP server intentionally exposes task-shaped tools rather than mirroring every REST endpoint one-to-one. The result: simpler inputs, clearer authentication expectations, stable tool names, and normalized error handling. A summary of the tool groups is available in Tool Surface.

How it works in short​

  1. An MCP client connects to the Etalasse MCP HTTP endpoint.
  2. The client completes the OAuth flow with an Etalasse Dashboard account — see OAuth Setup.
  3. Every tool request runs on behalf of the authenticated user, and the shop scope is resolved automatically from that session.
  4. The client calls tools/list to discover the available capabilities, then tools/call to perform actions — see Client Integration.
How Etalasse MCP works: the MCP client signs in through OAuth, then calls tools/list and tools/call on /mcp, which run as you in your shopDashboard sign-inOAuth, in the browsersame serverMCP clientClaude Code, Codex, ChatGPTEtalasse MCPPOST /mcpTool groupsShop dataCatalogOperationsStorefrontYour shopProducts, orderstools/listtools/callBearer tokenruns asyouShop scope comes fromthe session, not input123
(1) The client signs in once through OAuth with a Dashboard account, (2) then calls tools/list and tools/call on /mcp, (3) which run as you, in your shop.

(1) The client signs in once through OAuth with a Dashboard account, (2) then calls tools/list and tools/call on /mcp, (3) which run as you, in your shop.

Relationship to AI Chat and MCP Console​

The AI Chat feature in the Dashboard uses the same tool layer as this MCP server — that is why the in-product assistant can read your products, sales, and reports. The MCP Console in the Dashboard shows the connection health status and the list of active tools, which makes it useful for verifying your integration setup.

Both are described in the AI Workspace.