Install
Getting started

Why developers use Blamely

git blame tells you who committed a line. It doesn’t tell you whether that line was typed by a person or generated by an assistant. Blamely closes that gap with facts, not vibes.

For individual developers

  • Honest self-review. Before opening a PR, see the real AI vs human split of your diff. Decide where to add tests, comments, or a second pass.
  • Skill tracking. Watch the mix shift across weeks, projects, and branches — useful if you’re calibrating how heavily you lean on Copilot, Cursor, or Claude Code.
  • No more guessing. When a feature “felt” mostly AI-assisted, you can verify it instead of speculating in standup.

For code review

  • Targeted attention. Reviewers can prioritize files or commits with a high AI percentage for correctness, security, and style — without banning AI outright.
  • Better diffs. “This commit is 40% AI” is easier to discuss than a subjective claim. It complements git blame (who committed) with authorship in the editor (who wrote).
  • Regression hunting. When something breaks, attribution helps you check whether risky lines cluster in AI-heavy regions.

For teams and tech leads

  • Shared language. Teams use the same numbers, from the same tool, inside the same editor they already work in.
  • Capacity and tooling decisions. Coding time vs AI wait time and per-model breakdowns inform planning instead of replacing judgment.
  • Policy with evidence. Disclosure requirements or AI-use limits become measurable per commit instead of checkbox compliance.

For compliance and audit

  • Durable record. Optional git notes under refs/notes/blamely are tied to each commit and travel with the repo. See How it works.
  • Local first. Core attribution runs on your machine. Nothing is sent to Blamely or any third party. See Privacy & data.

What Blamely is not

  • It is not a replacement for code review, security scanning, or human judgment.
  • It does not score developers as “good” or “bad” — it attributes lines so you decide what norms and follow-ups make sense.

Something unclear or wrong? Tell us.