Codeball
AI code review that scores every PR for risk — auto-approving the safe ones, flagging the risky.
About
Codeball (by Codeball AI / Sturdy) is an AI code reviewer delivered as a GitHub Action (github.com/sturdy-dev/codeball-action). A deep-learning model trained on over a million pull requests assigns each PR a risk score from 0 to 1: safe contributions get auto-approved, risky ones get labeled for human attention, and teams can program custom review policies through modular, composable Actions. It supports 20+ languages (Python, TypeScript, Go, Rust, Java, and more), installs with sensible defaults, and measures DORA metrics so teams can track review quality over time.
The honest picture: the precision-tuned 'only approve what it's confident about' design makes it trustworthy as a first-pass filter rather than a rubber stamp — but that also means borderline PRs still need humans, so it reduces toil more than headcount. Pricing is a free tier plus paid plans (third-party listings vary, roughly $10–15/month — confirm current pricing on the site). It is GitHub-centric; GitLab/Bitbucket teams should look elsewhere.
What it can do
- ML risk scoring (0–1) on every pull request
- Auto-approval of safe PRs, labeling of risky ones
- Composable, programmable review policies via GitHub Actions
- 20+ language support with sensible defaults
- DORA metrics tracking for review quality
Setup
Add the Codeball GitHub Action to your repository workflow (one-click install from the marketplace). Configure risk thresholds and auto-approve/label policies in the action config.
Prerequisites
A GitHub repository; GitHub Actions enabled.
Use cases
Works with
Trust score — 6/10
Scored 6/10 (metrics tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 58; Trending score 60/100 from researcher assessment.
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.