--- title: "Announcing Our $3M Pre-Seed Round and StarSling Review Runners" description: "A $3M pre-seed from Bessemer Venture Partners, Y Combinator and five more funds to make shipping code agent-native, and Review Runners in private beta." date: 2026-09-22 url: https://starsling.dev/blog/pre-seed canonicalUrl: https://starsling.dev/blog/pre-seed --- # Announcing Our $3M Pre-Seed Round and StarSling Review Runners Today we're announcing that StarSling has raised a **$3 million pre-seed** round from Bessemer Venture Partners, Y Combinator, and several other amazing investors, and that we're launching [**StarSling Review Runners**](/products/review-runners). When we launched StarSling Runners this past April, the bottleneck had shifted from writing code to verifying it. That's still true today. Before code reaches production, it has to pass two gates: CI and code review. StarSling is now expanding to make both of those gates agent-native. ## First we built CI runners that get faster on their own [StarSling Runners](/products/runners) are a drop-in replacement for GitHub Actions' default `ubuntu-latest` runners. Jobs run on current-generation AMD EPYC hardware, roughly 30% faster than GitHub's, and switching is a label swap. Most CI services are static. Ours is [self-improving](/blog/ci-harness-v2-the-agent-loop-that-shipped-a-5-4x-speedup). StarSling's agents continuously analyze workflows, run logs and machine telemetry, then open pull requests that repair caching, speed up dependency installs, parallelize build steps and shard test suites. Every agent-authored change ships as a pull request your team reviews and merges, so the pipeline gets faster the longer you use it. You can see real examples of PRs that customers have shipped [here](/products/runners#ai-prs). Cost gets cut three ways. You get a lower per-minute rate than GitHub, faster machines that finish jobs in fewer minutes, and agents that remove minutes from the pipeline altogether. ## What customers measured [Image: Three bars to scale. Mastra's test suite falls from 29m 56s to 5m 06s, six times faster. Better Auth's end-to-end suite falls from 2m 22s to 1m 04s, twice as fast. Partcl's queue time falls from 9m 30s to 35s, sixteen times shorter.] [Partcl's](/customers/partcl) Rust tests got twice as fast within a day of migrating, off a pull request our agents opened. Across the platform we've now run over 2 million CI jobs, made CI up to 6x faster and up to 13x cheaper, and saved customers like [Mastra](/customers/mastra) and the [Better Auth](/customers/better-auth) team at Vercel over 28,000 hours in compute. > "At Mastra we move so fast that the bottleneck becomes reviews and CI. The runners are just handled, so nobody on my team thinks about CI infrastructure anymore." - Abhi Aiyer, Co-Founder and CTO at Mastra As of two weeks ago, tests that need a GPU get the same treatment on [StarSling GPU Runners](/products/gpu-runners). One of our early customers is already seeing a 14x reduction in queue time for a GPU runner! ## Now the second gate, code review Hosted review bots now let you customize their reviewer, but engineers still repeat themselves in every pull request. Review Runners let you build your own. We wanted the review itself to belong to the engineering team: the model, the skills, the scripts and the reviewer responsibilities, all versioned with the code and run in CI. You should have the freedom to customize your code reviews the same way you customize your coding agents. Coding agents load repository instructions, skills and tools that evolve with the codebase. Code review should work the same way. [StarSling Review Runners](/products/review-runners) run your model, your repo's review skills and scripts, and your team's reviewers as GitHub Actions jobs on every pull request, before a human opens it. [Image: A pull request comment from starsling carrying a high-severity finding on auth/session.ts line 142 and a medium one on api/routes/billing.ts line 64, signed by the security reviewer, running on your model and your key, one comment per pass.] Three things your team owns: - **Your model.** Bring your own provider key. The reviewer calls your model through your account, so the token bill stays with the provider you already use. - **Your review code.** Skills, scripts, instructions and reviewer configuration live in your repository, versioned alongside the code they review. Merge an update and future reviews use the new version. - **Your reviewers.** Define reviewers for things like security, APIs, tests, database changes or frontend architecture. Give each its own responsibilities and skills. For example, you could give a security reviewer your security scripts and an API reviewer your API conventions. The reviewer's skills and instructions load from the trusted base branch immediately before the model runs, so a pull request cannot rewrite the reviewer evaluating it. The job running the model holds a read-only token, and a separate step publishes the review. The diff, title, description and comments are treated as untrusted input. Each reviewer leaves one structured comment with a verdict, findings by severity, exact `path:line` references and a concrete fix. A new push cancels any pass still running on the previous commit, so the next completed review covers the latest code. Pay StarSling for the runner. Pay your model provider for the model. Reviews are billed by the minute like any other CI job. Run Review Runners on every pull request for your team's own review logic. Check out [the Review Runners product page](/products/review-runners) for more details. ## Why now, and what's next We started StarSling last year because coding agents got dramatically better, but shipping code hadn't. CI/CD, performance, and incidents are all still mostly manual. Just go look at the status page of your favorite model provider. DevOps isn't solved yet. We're starting where every team is currently bottlenecked: at the pull request. > "Every phase of the software development life cycle is being rebuilt to be agent-native, and CI is next. Daniel helped build the internal developer platform the 3,000+ engineers at Netflix relied on, and Yonas built StackShare into a community of more than a million developers. Their combined experience and understanding of developers at that scale is why we backed their vision to bring AI agents to DevOps infrastructure." - Elliott Robinson, Partner at Bessemer Venture Partners Additional investors in the round include Vermilion Cliffs Ventures, Precursor Ventures, Cervin Ventures, Outset Capital, and Transpose Platform. Angel investors in the round include David Cramer, Co-Founder of Sentry; Zach Holman, GitHub's second engineering hire; Idris Mokhtarzada, Co-Founder and CTO of Rocket Money; Kulveer Taggar, founder of Phosphor Capital; and engineering and product leaders from OpenAI, Anthropic, ClickHouse, Google, Microsoft, Meta, Coinbase, and Shopify. [Image: A StarSling card with a large $3M figure, the line 'To make shipping code agent-native', and the round's investor marks, Bessemer Venture Partners and Y Combinator above Vermilion Cliffs Ventures, Precursor Ventures, Cervin Ventures, Outset Capital and Transpose Platform.] We'll be using this funding for compute and to hire additional founding engineers in the San Francisco Bay Area. ## Get started Three ways for you to get up and running with StarSling: 1. **StarSling Runners** [Install the GitHub App](https://github.com/apps/starslingdev), swap one job to a StarSling label, and compare the run. Your first month includes 2,000 free minutes. 2. **Review Runners** [Sign up for early access](https://tally.so/r/obJPNN) and tell us which models you run and how much you spend on code review currently. 3. **GPU Runners** [Contact us to get started](https://tally.so/r/Y56E8v) and tell us which GPUs you need and at what scale. Happy Slinging! Yonas & Daniel