mcpField
D
doublespeed — Xcode as an API

doublespeed — Xcode as an API

Local
mac.doublespeed.aiDeveloper Tools0Updated 7w ago

About doublespeed — Xcode as an API

doublespeed provides coding agents with remote Macs for building, testing, and running iOS apps. Its tools match XcodeBuildMCP names, allowing workflows written for a local Mac to run without one.

Key features

  1. 01Discover Xcode projects and schemes
  2. 02Check project files and configuration
  3. 03Build and test iOS apps
  4. 04Run apps in live simulators
  5. 05Create unsigned archives and inspect jobs

Use cases

  1. 01Build iOS apps from Linux or a container
  2. 02Run XCTest or Swift Testing on remote Macs
  3. 03Preview an app in a live simulator
  4. 04Check project setup without compiling

MCP support

The MCP server exposes tools that run iOS build, test, archive, and simulator tasks on doublespeed's remote Mac fleet. Its tool names match XcodeBuildMCP.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
"mcpServers": {
"ai-doublespeed": {
"command": "npx",
"args": [
"-y",
"@doublespeed/mcp"
],
"env": {
"DS_API_KEY": "<DS_API_KEY>"
}
}
}
}
Run commandnpx -y @doublespeed/mcp
Environment variables to set
DS_API_KEYSecretdoublespeed API key (dsx_…) from https://mac.doublespeed.ai/dashboard

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
npm@doublespeed/mcp0.1.1

Tools

11

Saved tool list · updated 2w ago

ParameterTypeRequiredDescription
maxDepthnumberNo
workspaceRootstringNoDirectory to scan (default: cwd)

Details

From the project README

@doublespeed/mcp

An MCP server that gives a coding agent a Mac. Build, test, archive and run iOS apps — native, React Native, Expo — on the doublespeed Mac fleet from Linux, a container, CI, or a Mac without Xcode.

Tool names match XcodeBuildMCP (build_sim, test_sim, build_run_sim, list_sims, discover_projs, list_schemes, stop_app_sim, doctor) so prompts and skills written for a local Mac work unchanged when there is no Mac.

Install

Needs Node 20+. check_project works with no key at all; everything else needs a DS_API_KEY (free $100 of Mac time at https://mac.doublespeed.ai/dashboard → API keys).

Claude Code

claude mcp add doublespeed -e DS_API_KEY=dsx_... -- npx -y @doublespeed/mcp

Cursor / Windsurf / Codex / VS Code — add to the client's MCP config:

{ "mcpServers": { "doublespeed": { "command": "npx", "args": ["-y", "@doublespeed/mcp"], "env": { "DS_API_KEY": "dsx_..." } } } }

Tools

ToolWhat it doesWhere it runs
discover_projsfind .xcworkspace / .xcodeproj / Package.swiftlocal
list_schemesshared schemes of a project/workspacelocal
list_simssimulators, runtimes and Xcode versions on the fleetAPI
check_projectno key needed — schemes resolve, every Swift file parses, plists valid (lint, no compile)Mac
build_simcompile for the iOS Simulator; file:line:col errorsMac
test_simbuild + XCTest / Swift Testing; per-test resultsMac
build_run_simbuild, install, launch; returns a browser URL to a live simulatorMac
stop_app_simend a live sessionAPI
archive_appunsigned .xcarchive (Release)Mac
job_statusstatus, diagnostics, tests, log tail of any jobAPI
doctorkey, API reachability, fleet capabilitieslocal + API

Only files under workspaceRoot (default: the project's directory) are uploaded, content-addressed — after the first run only changed files move. A small app's test run takes ~20–40 s of Mac time and costs a few cents; nothing is billed while queued.

Not yet

Screenshots / accessibility tree of the running simulator, code signing and TestFlight upload. The live build_run_sim link is meant for a human to look at.

Agents without a Mac build, test and run iOS apps through mac.doublespeed.ai (npx -y @doublespeed/ds). Docs: https://mac.doublespeed.ai/docs.md

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.