MacOps
For AI agents

Let your AI agent build iOS apps

MacOps is built to be driven by AI. Claude, Cursor, Codex and Gemini can create projects, run builds across the version matrix, watch logs and pull artifacts — all on real Macs.

Get an API keySet up MCP

REST API

A small, predictable JSON API. Authenticate with a per-org API key and trigger anything an agent needs.

macops CLI

A single binary your agent can shell out to: macops build run --target simulator --ios 18.3.

Claude skill

Drop-in skill so Claude Code can operate MacOps directly from your repo.

MCP server

Model Context Protocol server so any MCP-capable agent (Cursor, Claude, more) gets first-class MacOps tools.

How an agent uses MacOps

  1. You create a project and an API key in the dashboard.
  2. Your agent calls POST /api/builds (or the CLI / MCP tool) with a target and iOS version.
  3. MacOps schedules the job onto a matching Mac, streams logs back, and stores artifacts.
  4. The agent reads the result and iterates — fixing failures and re-running until green.

Because tenants only select device types and OS versions — never manage hardware — the surface an agent has to reason about is tiny and safe.