AgentCloud
Log in Start building

an iPhone.

Create, operate, and inspect a hosted iOS Simulator through one OAuth-enabled MCP URL.

Start building See a recorded run
From $0.00033 / secondNo seats · usage credits included

Developer

For individual agent workflows

$49minimum
per month

$0.00050 per active second

  • About 27 simulator hours included
  • 1 concurrent session
  • 7-day evidence retention
Choose Developer

Team Recommended

For teams shipping with agents

$199minimum
per month

$0.00040 per active second

  • About 138 simulator hours included
  • 4 concurrent sessions
  • 30-day evidence retention
Choose Team

Scale

For high-volume agent infrastructure

$999minimum
per month

$0.00033 per active second

  • About 841 simulator hours included
  • 10 concurrent sessions
  • 90-day evidence retention
Choose Scale

Start with usage credit.
Pay for active seconds.

Your monthly minimum becomes simulator credit. When it runs out, usage continues at the same per-second rate—without seat fees or per-run charges.

Active time begins when the Simulator is ready and stops when the run is destroyed or reaches its inactivity timeout.

One MCP connection.
A complete iOS run.

Your agent gets a small, purpose-built tool surface—without raw shell access to the Mac worker.

Connect your MCP client
{
  "mcpServers": {
    "agentcloud": {
      "url": "YOUR_AGENTCLOUD_MCP_URL"
    }
  }
}

  1. Authorize once in the browser

    OAuth 2.1 and PKCE connect the MCP client to the right AgentCloud team. There is no agent secret to create or paste.

  2. Create a disposable Simulator

    Install a Simulator-compatible .app, launch it, open deep links, observe the screen, and act through semantic or coordinate input.

  3. Review the durable evidence

    The Simulator disappears at the end of the run. Its actions, screenshots, accessibility state, logs, and finalized recording remain inspectable.

Not just what failed.
What the agent actually saw.

Every action stays connected to the evidence available at that moment, so teams can reconstruct a run instead of guessing from an isolated test result.

ScreenVisual state before and after actions
UI treeAccessibility elements and semantic targets
LogsRecent Simulator and application output
TimelineActions aligned to screenshots and video
ArtifactsDownloadable evidence retained by the team

Small agent surface.
Explicit team boundary.

Simulator operations stay behind the application backend. Browser sessions, OAuth grants, runs, and artifacts are scoped to the team that created them.

Questions before the first run

The practical details for connecting an agent to a hosted iOS Simulator.

What is an iOS Simulator MCP server?

It exposes a focused set of iOS Simulator operations as Model Context Protocol tools. An MCP-compatible coding agent can create a run, install and launch an app, observe the screen and accessibility tree, act on the interface, read logs, and destroy the session.

Does the agent need to run on a Mac?

No. The iOS Simulator runs on AgentCloud’s Mac worker while the MCP client can connect remotely. Your app still needs to be built for the iOS Simulator runtime.

Can AgentCloud install an App Store IPA?

No. AgentCloud accepts Simulator-compatible .app bundles. A device-signed .ipa targets physical devices and is not interchangeable with a Simulator build.

What does AgentCloud retain?

Runs can retain actions, screenshots, accessibility snapshots, logs, typed values, artifacts, and full-run video. Because this evidence may contain sensitive data, teams should manage access and deletion deliberately.

How does per-second billing work?

Your monthly minimum is usage credit. Active Simulator time is totaled across sessions and billed at the plan’s exact per-second rate. Usage beyond the credit continues at the same rate.

Give the next agent run
somewhere real to happen.

Connect one MCP URL. Launch a disposable iOS Simulator. Keep the evidence.

Start building