Skip to main content

​
Overview

Connect Operator to Goose so its AI agent can call a dedicated, adaptive search service when it encounters confusing errors, missing docs, or unfamiliar integrations in your workflows. Instead of stopping at “I’m not sure”, Goose can delegate to Operator to scour the web, docs, and GitHub. This page walks you through configuring Operator as a Goose extension via the CLI or YAML, and gives a few practical ideas for how to use the combination in day-to-day automation and debugging tasks.

​
Configure Operator via the Goose CLI

The fastest way to add Operator is through the interactive Goose CLI: 
goose configure
In the wizard, choose “Add Extension”, then configure an HTTP extension that points to the hosted Operator MCP server. When prompted for the server URL, use https://mcp.operator.io/mcp and ensure your OPERATOR_API_KEY is set in your environment so Goose can send it as an Authorization: Bearer ... header. Each user supplies their own Operator API key. You can create and manage your key at operator.io. 
export OPERATOR_API_KEY="sk_your_api_key_here"

​
Configure Operator via Goose config file

You can also declare Operator directly in Goose’s YAML config, for example in ~/.config/goose/config.yaml: 
extensions:
  operator:
    type: http
    url: https://mcp.operator.io/mcp
    headers:
      Authorization: "Bearer ${OPERATOR_API_KEY}"
To use SSE instead of HTTP, change the url to https://mcp.operator.io/sse. You can also send the API key using an x-api-key header instead of Authorization. Once Operator is available as an extension, you can instruct Goose to use it for tasks like “look up the latest Prisma migrate docs”, “search GitHub for this test failure”, or “find an example of this library used with Next.js 15”.