Skip to main content
Beam runners use high-performance gaming machines and often run CI jobs 2–3× faster than GitHub-hosted runners. Beam starts a fresh Ubuntu 24.04 VM with Docker and Compose for each GitHub Actions job and deletes it when the job ends. Move an existing Linux job to Beam by changing its runs-on label.

Set up a runner

  1. Open Virtual Machines → GitHub Runners in the Beam dashboard and select Connect GitHub.
  2. Install the Beam GitHub App on your personal account or an organization you own. Grant it access to the repositories you want to run on Beam. If the app is already installed, select that account in Beam.
  3. Set runs-on to a single Beam label in your workflow. Run the workflow and check the job in the GitHub Runners tab.

Runner sizes and pricing

Beam bills runner compute by the second while the VM is running. Each size includes 50 GiB of writable disk. The table uses published Beam Performance tier rates, expressed per minute. At 2 vCPU, Beam’s Linux x64 rate is 25% lower than Blacksmith ($0.004/min) and 50% lower than GitHub-hosted runners ($0.006/min). This compares paid list rates before included usage. See Blacksmith pricing and GitHub Actions runner pricing.

Runner capacity

Beam starts runners on demand. Each workspace can have 100 concurrent runners by default, including runners that are starting. The workspace’s CPU limit and available compute also affect how many jobs can run at once. Jobs stay queued in GitHub until a runner is available. The GitHub Runners tab shows how many runners are in use, your workspace’s runner limit, and why jobs are waiting.

Enhanced cache

Enhanced cache uses Beam’s content-addressed storage to automatically reuse build artifacts and downloaded dependencies across jobs in the same repository. It covers Go build outputs and modules, npm packages, and pip downloads. It’s on by default and requires no workflow changes. Successful jobs from pushes to the default branch or scheduled runs on that branch populate the cache. Manage it under GitHub Runners → Settings.