Source Trace tags every line to the person, agent, and model. Tracks how much AI code was generated vs committed in Git. Finds which models work best in your code. Shows model names in Git history.
Installation
Launch VS Code Quick Open (Ctrl+P), paste the following command, and press enter.
AI tracks tokens. Git tracks commits. Neither tells the full story: who exactly wrote that line - you or AI? How much code survives? Which model you used last month?
Source Trace solves this by tagging every line to the person, agent, and model. It tracks how much AI code was generated vs committed in Git, and shows which models work best in your codebase.
Getting Started
Install the Source Trace extension
Open the Source Trace panel and click "Log In"
Start coding - Source Trace automatically tracks your AI coding
Authentication for the extension uses OAuth login with Google in your browser, then redirects back to VS Code.
Key Features
AI Blame: see exactly what code was generated by AI - even when the commit is authored by you or your teammate. Find which agents and models were used, months later.
Track Code Churn: visualize daily coding stats across AI and manual edits - even code which was written but didn't survive. Head to https://app.srctrace.com to see your data
Compare Models: find which AI agents and models produce code that survives review and work best in your codebase.
Privacy
We do not sell your data. We are not affiliated with AI model or cloud providers.
Like in best-in-class tools, plain text code does not leave your machine.
AI Blame uses your Git for storage.
Repository URLs, file names, and commit data are hashed with a one-way function, so we can deduplicate data and power your analytics dashboard.
More Information
Supported AI Tools: Codex, Claude, Copilot, Gemini, Kilo Code, RooCode. More coming soon!