Synthetic code. Real confidence.
Surface AI-authored code for extra review and safer releases.
Proof in every PR
Instant, line-level detection of AI-generated code with confidence scores—posted as GitHub Check annotations so reviewers can act before merge.
- ✓ Line-level annotations with confidence scores
- ✓ Severity levels: notice · warning · failure
- ✓ Checks tab summary + inline PR details
- ✓ Mark as required to block risky merges
Line-level provenance inside GitHub
When a pull request opens, reopens, or syncs, our webhook fires a Check Run that scans the diff and annotates each file with per-line detections. You’ll see notice / warning / failure levels and a confidence score for every flagged range—so reviewers know exactly where to focus and why.
Make the check required to enforce “no provenance, no pass.” Tune thresholds, exclude paths, and auto-route to owners. Built for CTOs, security teams, and maintainers who need visibility, policy enforcement, and audit-ready evidence without changing CI. Your workflow stays fast; surprises don’t ship.
Install our Githup Application
One-click install
Add to selected repos and start annotating PRs immediately—no CI changes required.
Policy gates that fit you
Set notice/warn/fail thresholds, filter paths, and auto-request reviewers.
Privacy-first
Detection runs via GitHub Checks; your code stays in your GitHub environment.
Audit-ready evidence
Confidence scores and line ranges create a clear trail for compliance and reviews.
We detect AI-authored code in the following languages:
