Founding price locked for the lifetime of the subscription.
- M4 Pro · 6 vCPU · 22GB
- 1 SCU · 1 concurrent job
- Lifetime price · max 2 founding SCUs
Fast, ephemeral M4 Pro runners for GitHub Actions. Keep every workflow, change one line, and pay a flat monthly price per concurrent slot.
Set up jobFresh Tart VM · M4 Pro
Checkout repositoryactions/checkout@v4
Compile macOS appXcode 16.4 · Swift 6.1
Destroy runnerSecrets and workspace erased
# Before
- runs-on: macos-latest
# After
+ runs-on: manzanita-standard
One slot runs one job at a time. Queue the rest in GitHub and run as many CI minutes as your team needs.
Founding price locked for the lifetime of the subscription.
The default runner for everyday Swift and Xcode workflows.
A full M4 Pro host profile for large builds and test suites.
Simple capacity: 1 SCU is half an M4 Pro: 6 vCPU, 22GB, and one concurrent job. Founding and Standard each use 1 SCU; XL uses 2 SCUs and receives the whole host.
Free trial: one Standard SCU for 14 days or 300 build minutes. Fair use: above 15,000 minutes per SCU in 30 days starts a capacity review, never a surprise charge.
Manzanita replaces only the slow machine underneath your existing job.
Grant access only to the repositories that should use Manzanita.
Keep your actions, secrets, cache keys, artifacts, and release flow.
Every job starts in a prewarmed Tart VM and the VM is destroyed afterward.
Your source, secrets, logs, checks, and artifacts remain in GitHub. Manzanita receives a short-lived job assignment, runs it inside one disposable macOS VM, then destroys the workspace and credentials.
Read the security model →There is no proprietary workflow language and no second place to debug your builds.
jobs:
test:
runs-on: manzanita-standard
steps:
- uses: actions/checkout@v4
- run: swift test
Fresh macOS Sequoia runner · Xcode 16.4
Dedicated Apple Silicon, canary-tested images, and fleet-wide health monitoring. Capacity is provisioned ahead of demand across isolated runner pools.
Connect your GitHub organization, keep your existing workflows, and start with a clean Standard runner.