Optional integration · Developer guide

Already live in your AI client? Connect a copilot over MCP.

Every My Copilot AI product works fully on its own, through the assistant built into it. This page is for the optional path: if you would rather ask from inside ChatGPT, Claude, Gemini, Cursor or VS Code, Model Context Protocol brings the same copilot to you.

Stdio / SSE

Transport Agnostic

Our servers run via standard process input/output (stdio) locally on your workstation, or via encrypted Server-Sent Events (SSE) for remote agents.

Revocable Keys

Stateless Auth

Studio API tokens provide cryptographically scoped access. Each MCP session is stateless and isolated; token revocation is instantaneous from your web dashboard.

< 120ms Latency

High Token Density

Tools return concise markdown abstractions specifically formatted for LLM context windows, maintaining <150ms execution latencies without cluttering chat history.

Developer MCP Setup

Select your runtime to view the standard JSON declaration.

{
  "mcpServers": {
    "money-copilot": {
      "command": "npx",
      "args": ["-y", "@copilotai/money-mcp@latest"],
      "env": {
        "MONEY_API_KEY": "copilot_live_sec_your_key"
      }
    },
    "frame-copilot": {
      "command": "npx",
      "args": ["-y", "frame-copilot-ai"]
    },
    "billable-copilot": {
      "command": "npx",
      "args": ["-y", "@copilotai/billable-mcp@latest"],
      "env": {
        "BILLABLE_API_KEY": "copilot_live_sec_your_key"
      }
    }
  }
}

Per-product integration docs

You never need these to use a product — its own copilot is already built in. They are here for when you would rather drive it from your own AI client: