Related Tools
Several tools address the challenge of tracking AI contributions to codebases. This document compares different approaches and explains where ai-blame fits in the landscape.
The Provenance Problem
As AI agents become routine collaborators in software development and knowledge curation, we face a new challenge: attribution. Traditional version control tells us who committed changes, but in an AI-assisted workflow, the human commits changes that an AI actually wrote.
Different tools take different approaches to solving this.
git-ai
Website: usegitai.com
git-ai takes a git-native approach to AI attribution. Rather than modifying files, it extends git's metadata system.
Comparison with ai-blame
| Aspect | git-ai | ai-blame |
|---|---|---|
| Granularity | Line-level | File-level (plus ai-blame blame for a best-effort view) |
| Storage | Git notes (.git/) |
Embedded in files |
| Timing | Real-time during coding | Post-hoc extraction |
| Portability | Via git clone | Files carry their history |
| Use case | Development workflows | Knowledge bases, structured data |