Run macOS or Linux VMs on Apple Silicon for automated build and test pipelines
Share and distribute VM images through OCI-compatible container registries
Replace or supplement GitHub Actions runners with faster cheaper Apple Silicon runners
Automate VM image creation using the Packer plugin
Requires Apple Silicon Mac with macOS 13 Ventura or later. Install via Homebrew, then pull a base image (~25GB download). Three-command quickstart available in README.
Tart is a toolset for building, running, and managing virtual machines on Apple Silicon Macs. It supports both macOS and Linux VMs and is built primarily for use in automated build and test pipelines, though it can be used for other automation tasks as well. It requires macOS 13.0 (Ventura) or later on an Apple Silicon device. Virtual machines in Tart run using Apple's own virtualization framework, which gives them performance close to running natively on the hardware. You can store and share VM images through any container registry that uses the OCI standard, the same kind used for Docker images. There is also a plugin for a tool called Packer that can automate the creation of new VM images. Tart is designed to integrate with CI systems, which are tools that automatically build and test software when code changes. The project powers a commercial service called Cirrus Runners, which provides faster and cheaper alternatives to the default runners offered by GitHub Actions. Several well-known companies are listed as users of Tart in their internal setups, including Atlassian, Figma, and Expo. Installing and running a VM is straightforward. The README shows a three-command sequence: install Tart via Homebrew, download a base macOS image from a container registry, and run it. The first download is around 25 GB. Full documentation is available at tart.run, and there is a GitHub Discussions section for questions. The project is also available through AWS Marketplace.
← cirruslabs on gitmyhub — every repo by this author, as a profile.
Verify against the repo before relying on details.