Buildermark

Buildermark is a desktop application that measures how much of your code is written by AI coding agents. It works by automatically importing your chat history from tools like Claude Code, Codex, and Gemini, then matching the code diffs from those conversations with the diffs in your git commits. This lets you see the exact percentage of each commit that was authored by an AI, without needing to change your workflow or rely on agent-specific hooks.

The tool stands out because it's entirely local-first. A Go server runs on your machine, and all your data stays there—there's no cloud sync, no accounts, and no analytics. It supports a range of agents through both desktop apps and browser extensions, and its matching algorithm is designed to handle formatting changes and line reordering. You also get native desktop notifications that show the agent attribution percentage the moment you commit.

Buildermark is most useful for developers and engineering teams who are actively using AI coding assistants and want to understand their impact. It helps answer practical questions: How much is AI actually writing? Which agent performs best on your specific codebase? For tech leads, it provides data on how AI adoption is changing the development process across a project or team.