---
title: "StarSling Runners: fast GitHub Actions runners"
description: "GitHub Actions runners on 5th Gen AMD EPYC, from 2 to 64 vCPU, at 21% to 33% less per minute, plus AI agents that keep opening PRs to make your runs faster."
url: https://starsling.dev/products/runners
canonicalUrl: https://starsling.dev/products/runners
---

# StarSling Runners

GitHub Actions runners on 5th Gen AMD EPYC, from 2 to 64 vCPU, at 21% to 33% less per minute, plus AI agents that keep opening PRs to make your runs faster.

## What makes it faster

5th Gen AMD EPYC, and the agent pull requests that shard suites, fix caching and cut critical path.

| Customer | Speedup | Measured | Change |
| --- | --- | --- | --- |
| [Partcl](https://starsling.dev/customers/partcl) | 16x | Shorter queue time | 9.5 min -> 35 s |
| [Mastra](https://starsling.dev/customers/mastra) | 6x | Faster test suite | 29m 56s -> 5m 06s wall p95 |
| [Better Auth](https://starsling.dev/customers/better-auth) | 2x | Faster E2E | 2m 22s -> 1m 04s |

Read the case studies: https://starsling.dev/customers

## Real PRs from StarSling agents

Pull Requests merged by customers. StarSling agents analyze and optimize your CI automatically.

- **Mastra: ~3.4 minutes saved per run** - Kitchen-sink E2E benchmark: 467s -> about 260s for the slowest shard across five warm-cache runs. [ci: shard E2E kitchen-sink across 3 parallel jobs](https://github.com/mastra-ai/mastra/pull/15888)
- **Better Auth: ~6,600 CI minutes/month** - Estimated time saved by caching Chromium and removing browser installs from six adapter jobs. [chore(ci): optimize Playwright browser installs in E2E](https://github.com/better-auth/better-auth/pull/8073)
- **Better Auth: Local build caching restored** - Enabled local and remote Turborepo caching across three CI workflows. [chore: fix turbo cache configuration in ci](https://github.com/better-auth/better-auth/pull/7950)
- **Mastra: 160s -> 16s of fixed waits** - Chroma's 80 indexing waits dropped from 2,000ms to 200ms each. The PR estimates 2.4 minutes saved per job. [test(chroma): reduce fixed 2000ms waitForIndexing sleep to 200ms and add Docker healthcheck](https://github.com/mastra-ai/mastra/pull/13866)
- **Mastra: 88% less checkout data** - Full-history clone transfer: 1,572 MB -> 189 MB. The PR estimates 445 runner-hours saved per month. [perf(ci): fetch full-history checkouts with --filter=blob:none](https://github.com/mastra-ai/mastra/pull/21298)
- **Mastra: Publishing tokens scoped to publish jobs** - Removed OIDC token access from the nonpublishing job; the three publishing jobs retain it. [ci(security): scope id-token: write to publish jobs only](https://github.com/mastra-ai/mastra/pull/16509)

## The lineup

| Label | Size | StarSling | GitHub | You save |
| --- | --- | --- | --- | --- |
| `starsling-ubuntu-24.04-2` | 2 vCPU / 8 GB | $0.004/min | $0.006/min | 33% |
| `starsling-ubuntu-24.04` | 4 vCPU / 16 GB | $0.008/min | $0.012/min | 33% |
| `starsling-ubuntu-24.04-8` | 8 vCPU / 32 GB | $0.016/min | $0.022/min | 27% |
| `starsling-ubuntu-24.04-16` | 16 vCPU / 64 GB | $0.032/min | $0.042/min | 24% |
| `starsling-ubuntu-24.04-32` | 32 vCPU / 128 GB | $0.064/min | $0.082/min | 22% |
| `starsling-ubuntu-24.04-64` | 64 vCPU / 256 GB | $0.128/min | $0.162/min | 21% |

Every size runs Ubuntu 24.04 on 5th Gen AMD EPYC. `ubuntu-latest` and
`ubuntu-24.04` both map to `starsling-ubuntu-24.04`. Billing runs from job
start to finish, rounded up to the nearest minute; queue time is not billed. GitHub
prices are its Linux x64 larger-runner list prices.

## Use it in a workflow

Before:

```yaml
jobs:
  build:
    runs-on: ubuntu-latest
```

After:

```yaml
jobs:
  build:
    runs-on: starsling-ubuntu-24.04
```

## Other StarSling runners

- [StarSling GPU Runners](https://starsling.dev/products/gpu-runners): Dedicated NVIDIA GPUs for the tests that need one

Label reference: https://docs.starsling.dev/configuration/label-reference

Latest labels and pricing: https://docs.starsling.dev/runners/compute-sizing