CI runs the workflow files in your repositories on every push, pull request, merge queue build or manual start: native-format workflows and GitHub Actions workflows as they are, on machines the platform creates for a run, on runner pools it starts in your organization, or on Linux machines and Macs you register, with environments, variables, OIDC tokens and required checks.
CI runs the workflow files in your repositories on every push, pull request, tag or manual start, on machines the platform creates or on machines you own.
Add a workflow to a repository, watch its first run and read its log.
Where workflow files are read from, what starts a run, where it executes, and the statuses it goes through.
Write a workflow file, check it with the same parser a push runs, and commit it to the repository.
Push a repository with its .github/workflows folder and its workflows run here as they are, with uses actions, matrices, caches and artifacts.
Issue a one-time registration token and register a machine you own as a runner, with GitHub's runner on Linux or by hand on any system.
Have the platform start machines in your organization for GitHub-syntax jobs with the pool's labels, one job per machine, and stop them when the work is done.
Who may read, start, cancel and configure what (IAM), what is audited, and how runs are kept from reaching machines and secrets they should not.
Where to see what runs, what waits and why, how your runners and pools are doing, and what CI used.
Which machine executes a run, what credentials it holds and for how long, what the runner software on your own machines does, and what happens when something fails.
User Guide
31 pages: concepts, tasks, security, quotas, troubleshooting.
API Reference
37 operations, with their parameters and IAM actions.
CLI Command Reference
37 commands, one per API operation, with their arguments and flags.
Terraform resources
1 resources and data sources of the provider.
Document history
What changed in these docs.