StarSling launches self-improving CI tool for AI agent-heavy repos
Y Combinator-backed startup StarSling has launched a code review service that runs as a GitHub Actions job on its own dedicated runners. The service, called Review Runner, uses 4 vCPU/16GB machines priced at $0.004 per minute, with new accounts billed per second and no minimum charge, while queue time is never billed.
GoKawiil's interpretation of the reporting above, not reported fact.
The pricing structure and unlimited concurrency suggest StarSling is targeting teams running many parallel AI-agent-generated pull requests, where traditional per-minute CI billing could be wasteful. By separating compute billing from model token costs (paid directly to providers), the company avoids taking a cut of AI usage, which could make it more attractive to developers already paying for LLM access elsewhere.
- StarSling's Review Runner is a GitHub Actions-based automated code review tool aimed at agent-heavy repositories
- New subscriptions bill per-second with no minimum, while legacy accounts still use per-minute billing until migrated
- Model token costs are billed separately by the user's own AI provider, not by StarSling
Source: starsling.dev, 2026-11-17
Published there as: “Launch HN: StarSling (YC P25) – Self-improving CI for agent-volume repos”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.