MCP Integration

What is MCP?

The Model Context Protocol (MCP) is an open standard that lets AI assistants interact with external tools and APIs. AI Studio exposes an MCP server that allows you to build adaptive models, manage datasets, run inference, and more — all through natural language.

Quick Setup

1. Get an API key from the API Keys page (requires login).
2. Claude Code — run this in your terminal:
   Copy claude mcp add --transport http esmart https://ai.esmartsystems.com/mcp --header "x-api-key: YOUR_API_KEY"
3. Claude Desktop / Cursor / Windsurf — add to your MCP config:
   Copy

   {
     "mcpServers": {
       "esmart-ai": {
         "url": "https://ai.esmartsystems.com/mcp",
         "headers": {
           "X-API-Key": "YOUR_API_KEY"
         }
       }
     }
   }

4. Start a conversation and ask your AI assistant to build a model, import images, or run inference.

Capabilities

Once connected, the MCP endpoint automatically exposes all available tools and their descriptions to your AI assistant. No manual configuration needed — the assistant discovers what it can do and how to use each operation through the protocol itself.

This includes dataset management, adaptive model building, few-shot classification, cloud storage import, and more. Destructive operations (delete, update) are intentionally excluded to prevent accidental data loss.