XR Build 16 · Dedicated physical Mac mini
NewMac mini M4 Pro 64GB / 2TB now available
A dedicated Apple Silicon Mac mini—not a virtual machine. Choose daily, weekly, monthly, or quarterly billing. SSH and VNC credentials are provided after checkout, so you can use it as your iOS build Mac right away.
6 locations
$108.3/month · XR Build 16
Workspace
MyApp.xcworkspace
GroupsArtifacts
$ xcodebuild -workspace MyApp.xcworkspace -scheme MyApp -configuration Release archive
MyApp.ipacom.example.myapp · 2.4.1 (318)
Archive SucceededSigned6m12s
Illustrative build workflow, not a performance benchmark. Actual build times depend on your project, caches, dependency downloads, and concurrent workloads.
Inspect
XR Build Pro 64
NodeLogs
› sudo xcode-select -s /Applications/Xcode_16.4.app
Every plan gives you a complete physical node, and all three configurations are available at all 6 locations. Start with a day to test your project or build continuously month to month. Weekly and quarterly billing, plus storage upgrades, are listed on the plans page. Availability is confirmed live in the portal.
M4 / 16GB / 256GB SSD
$21.7/day
$108.3/month
Get an app’s Archive and export workflow up and running. A good fit for solo development and everyday builds run sequentially. Clear out DerivedData and old artifacts regularly to leave room for source files and dependency caches.
M4 / 24GB / 512GB SSD
$40.8/day
$204.1/month
More headroom for multiple targets, widgets, extensions, and UI tests. Suited to team pipelines that keep several dependency caches around. Set concurrency based on your project’s actual memory use; thread count is not the same as the number of tasks you can run in parallel.
M4 Pro / 64GB / 2TB SSD
$59.5/day
$297.3/month
64GB of unified memory for simulator test matrices, larger caches, and MLX experiments. Benchmark one task before increasing concurrency. Schedule model training and release builds at different times to keep workloads from competing for resources on the same node.
Payment is accepted only via USDT-TRC20 or Visa / Mastercard / Amex (through Stripe). All charges are in USD. Available payment options are confirmed by the checkout service.
These are practical workflow examples, not customer case studies or performance guarantees. The commands show where to start; prepare your project dependencies, signing materials, and required software licenses before running them.
After a commit, use launchd or a runner to check out the specified revision, install dependencies, run tests, archive, and export in sequence. Upload to TestFlight with Fastlane, then review build logs and test results the next day before deciding whether to distribute to testers.
Keep your Scheme, build settings, and export options consistent, and store signing materials in the node’s own Keychain. Verify signing once in the graphical interface before switching to unattended builds.
xcodebuild -scheme MyApp archive -archivePath build/MyApp.xcarchive
Register a GitLab or GitHub self-hosted runner on the node, then route jobs by project, toolchain, and environment labels. Once a pull request is merged, run Fastlane gym to export signed artifacts for the main app and its extensions, and retain logs so failed jobs can be retried.
A dedicated node prevents other tenants from competing for its resources, but your team still needs to manage its own job queue. Set sensible runner concurrency and isolate working directories. Track certificate changes, Xcode switches, and cache cleanup in version-controlled scripts.
fastlane gym --scheme MyApp --export_method app-store
Load a model with MLX, prepare your data, and run LoRA fine-tuning. Depending on model support, you can then use coremltools to convert it to Core ML. Add the result to your iOS project and check operator compatibility, inference memory use, and output quality.
Unified memory is shared by the model, training state, and macOS, so 64GB is not all available to the model. Start with a small batch to check memory use before increasing context length. Schedule training and packaging at different times rather than assuming both should run at once.
python -m mlx_lm.lora --model ./base --train --data ./data
Connect to the macOS desktop over VNC and schedule renders in a licensed Logic or Final Cut setup. For command-line jobs, encode with VideoToolbox and retrieve the finished files over SSH or rsync.
Upload a representative sample first to check plugins, fonts, codecs, and output settings. Transfer times for large files depend on your local bandwidth and network route, so consider scheduling bulk uploads separately from rendering.
ffmpeg -i in.mov -c:v hevc_videotoolbox -b:v 12M out.mp4
Monthly uptime commitment
If monthly uptime falls below 99.9%, qualifying downtime is credited as additional rental time. This status bar illustrates how daily records are displayed; it does not indicate that this site has accumulated 90 days of uptime history or claim a particular actual uptime rate.
If you have connection issues, first check whether the cause is your local network, login credentials, or node status. When submitting a support ticket, include the instance, when the issue began, your time zone, and any error messages. The terms of service define the limits for user actions and force majeure.
All three configurations are available at all 6 locations. Choose a city to open the order form with that location preselected. Availability is confirmed live in the portal.
For remote desktop work, consider where your team is based and the network route. For automated builds, factor in the transfer of source code, dependencies, and artifacts. If your team spans time zones, use UTC for task timestamps to keep logs and schedules aligned.
US locations display standard-time offsets; local offsets change when daylight saving time applies. Time zones help with scheduling but do not guarantee network latency from every access location.
Your build environment is not shared with other tenants. You manage your own source code, dependency caches, simulators, and signing chain. Turn the setup into scripts to make each archive easier to reproduce.
No overselling or partitioning. Your workflows get the CPU, GPU, and unified memory, with no other tenants’ build jobs running on the machine. You still need to manage your own concurrency to prevent multiple archives from competing for disk and memory.
macOS runs directly on an Apple Silicon physical node. Builds, simulators, and Metal workloads use the machine’s own resources. You’ll still need to test your apps, dependencies, and tools for compatibility with your project; the hardware model alone is no substitute for testing.
You get a full macOS administrator account, with complete access to both the graphical interface and command line. Switch toolchains with xcode-select -s. In CI, use DEVELOPER_DIR to specify a version for an individual job without changing the global settings for other jobs.
Choose a configuration, location, billing period, and add-ons, then review the USD price and pay.
Your dedicated physical machine is delivered with macOS and Xcode installed. Check the instance details and toolchain.
Get SSH and VNC credentials from the portal, run a test build, and then register your runner.
Plan your Xcode version, signing materials, backups, and renewals so you know what to check during your first build.
Yes. Install each version in a separate application directory. Check your project’s dependencies and SDK requirements, then switch versions with sudo xcode-select -s. In CI, set DEVELOPER_DIR for each job where possible, and log xcodebuild -version to make it easier to troubleshoot after someone changes the global version.
Import the certificate and private key into the node’s Keychain, then install a provisioning profile that matches your Bundle ID, team, and export method. Verify signing once in the graphical interface before configuring Keychain unlocking and access for command-line jobs. Never commit private keys, passwords, or runner tokens to your source repository.
A reinstall erases data on the system disk. Before proceeding, back up your source code, signing materials, configuration, and any required caches somewhere off the node, and check that you can restore them. The +1TB / +2TB SSD options expand storage; they are not separate backups. Submit a support ticket to confirm how the relevant disks will be handled before reinstalling.
Sign in to the portal to check your rental period and billing details, then follow the prompts to renew. The portal will confirm your payment status. Payment is accepted only via USDT-TRC20 or Visa / Mastercard / Amex (through Stripe), in USD. Before you stop using the node, retrieve your artifacts and data, then revoke deployment keys and runner credentials.
Explore practical guides on toolchains, dependency caches, and runner setup, then add the steps that suit your project to your own pipeline.
Separate ZIP imports into entry validation, bounded writes and commit. Use Swift negative tests to catch traversal, then verify links and extraction quotas.
DevOps & CI/CDStart with a small XCTest fixture for grouped edits and redo. Extend it to new edits, document changes and stale asynchronous results.
DevOps & CI/CDExtract monetary rules into a Swift package and test decimal parsing, negative ties, and rounding boundaries without launching a simulator.
New articles are in the works. In the meantime, browse the blog.
From $21.7/day · 6 locations · SSH and VNC credentials included
Start with a representative project and verify dependencies, signing, and export before deciding on concurrency and a longer billing period. Node availability is confirmed live in the portal.