What's Actually Happening
A developer documented running six automation workflows - market scanning, deal alerts, lead monitoring, and data aggregation - on GitHub Actions' free tier. The setup processes approximately 1,400 minutes monthly across scheduled Python scripts, staying within the 2,000-minute public repository limit.
The architecture stores results as CSV files committed to Git (eliminating database costs), triggers workflows via cron scheduling (minimum five-minute intervals), and sends notifications through ntfy.sh's free tier.
The Trade-Offs
This works for specific use cases. GitHub Actions' free tier provides unlimited minutes for public repositories and standard runners, but comes with constraints: five-minute minimum scheduling intervals, 15-minute variance in cron execution, six-hour maximum job runtime, and public code visibility.
For private repositories, limits drop sharply: 2,000 minutes monthly on free plans, 3,000 on Pro ($4/month), 50,000 on Enterprise. Overages bill automatically unless spending limits are configured. Larger or GPU runners incur costs regardless of repository visibility.
The developer reports using this for storage arbitrage opportunities and tool rental, claiming $800 first-month profit. Worth noting: the workflows are being sold as templates ($79 for the full bundle), which changes the economics substantially.
What This Means in Practice
For enterprise teams evaluating workflow automation costs, the math gets interesting at scale. GitHub Actions undercuts Zapier ($240/year) and n8n Cloud ($240/year) for light automation, but self-hosted runners remain the only truly unlimited option - always free, regardless of compute volume.
The real question is whether public code requirements and scheduling limitations fit your use case. For market scanning or periodic data aggregation with non-sensitive logic, this approach works. For real-time operations, proprietary workflows, or high-frequency tasks, it doesn't.
Alternatives worth considering: GitLab self-hosted runners (comparable costs, different ecosystem), Gitea/Forgejo for fully self-hosted Git plus automation, or self-hosted GitHub runners on minimal hardware (Raspberry Pi setups documented extensively). The pattern here - scheduled scripts committing to Git - ports to any CI/CD platform.
History suggests free tier automation strategies work until they don't. The constraint that breaks first is usually scheduling precision or minute consumption, not storage.