StarSling Runners
AI-native CI for GitHub Actions, up to 6x faster·starsling.dev
StarSling Runners are AI-native GitHub Actions runners that serve as a drop-in replacement for ubuntu-latest and ubuntu-24.04, running on 5th Gen AMD EPYC hardware. AI agents continuously analyze workflows, run logs, and machine telemetry to open pull requests that optimize caching, dependency installation, test configuration, and workflow structure. Runners are billed per minute of usage with unlimited concurrency, no seat fees, and queue time that is never billed. Each job runs in a single-use, hardware-isolated microVM for isolation.
What it's for
Features 19
- 5th Gen AMD EPYC processors
- 99.99% SLA
- AI agents that continuously analyze workflows and open optimization pull requests
- Build step optimization (parallelization, removal of redundant/slow steps)
- Caching optimization (detects missing or misconfigured caches)
- Compatible with actions/cache with zero config
- Dependency installation optimization (e.g. frozen lockfiles, parallel installs)
- Drop-in replacement for ubuntu-latest and ubuntu-24.04 runner labels
- No seat fees
- One-click GitHub App installation
- Open-source ci-score skill grading workflow configuration against an 11-check rubric
- Open-source ci-speedup skill auditing workflows against ~73 CI optimization patterns
- Per-minute usage-based pricing
- Queue time never billed
- Secrets passed directly from GitHub to the job, never stored by StarSling
- Single-use, hardware-isolated microVM per job
- Test optimization (misconfigured test jobs, test database interactions, test sharding)
- Unlimited concurrency
- Workflow structure optimization (job splitting, matrix strategies, dependency ordering)
At a glance
Integrations
Resources
Pricing
Pricing is usage-based per minute of runner time rather than a flat monthly fee. Runner rates range from $0.004/min (2 vCPU/8GB) to $0.128/min (64 vCPU/256GB); the standard 4 vCPU/16GB runner is $0.008/min. New accounts get 2,000 free minutes in the first month. A page element referencing an 'Alum Discount' appeared near the Enterprise tier but no further detail was given.
AI-powered optimization PRs are only available to customers on paid plans and are not enabled by default for new accounts.
2,000 free minutes in the first month
Queue time never billed, Unlimited concurrency, 99.99% SLA
Billed per minute of runner usage; queue time never billed
Volume discounts available; contact a founder for a quote.
Custom minute allocation
Volume discounts available on the Enterprise plan.