ShoalStack is now
With the platform201 MCP server, Claude, ChatGPT, Cursor, VS Code, and other MCP clients can work with your projects, environments, Blueprints, and infrastructure graph.
After you connect, your assistant acts as you, inside the spaces your account can access.
It works with the same projects, environments, and graph that you see in platform201.
Ask for the list of Blueprints, the shared templates
on the platform. Your assistant reads their graphs
and descriptions before you pick one.
Ask your assistant to create a project of your own
and add an environment to it. It works inside the
space you signed in to.
Give your assistant the variables that your services
need. It adds each one to the environment, so the
nodes can refer to it by name.
Ask what an environment contains. Your assistant
reads its nodes, connections, and settings, and
answers from the current graph.
Most developers already work with an AI assistant in an editor, a terminal, or a chat app. Connect it to platform201, and infrastructure tasks stay in the same window as your code.
Cursor, Windsurf, Cline, and VS Code each connect with one config file.
Claude Desktop, ChatGPT, and Grok, each connect through a custom connector.
Claude Code, Codex, and Gemini CLI each add the server in one step.
Each client connects in a similar way. This sample uses Cursor: you add the server, sign in once, check the connection, and then give your assistant its first request.
Create a Cursor MCP config file for your project, and add the platform201 MCP server URL path to it. Then enable the server in Cursor.
Select the login prompt. Your browser opens, and you sign in
to platform201 and approve access. Cursor stores and
refreshes the token.
Ask Cursor: “List the available platform201 Blueprints.” If you get a list of Blueprints in the reply, it means that the connection works.
Ask Cursor to create a project and add an environment to it. The new project will appear in your platform201 space, ready for services.
An outside assistant is one way to use AI with platform201. You can also use the built-in assistant, review its work on the graph, or run an MCP server of your own.
Create your platform201 space, add the server to the AI tool you already use, and sign in once. Then ask your assistant for the Blueprints, a project, or an environment.
The Model Context Protocol is an open standard that lets AI assistants use outside tools and data. A service runs an MCP server that offers a set of tools, and an AI client calls those tools for the user. With MCP, one integration works across many assistants.
A client needs three details to connect to a remote MCP server:
• The server URL
• The transport, such as Streamable HTTP
• The sign-in method, such as OAuth or an API key
The platform201 MCP server uses Streamable HTTP and OAuth, so AI clients can work with your projects, environments, Blueprints, and infrastructure graph.