Skip to content

Selvedge vs. Git AI

Git AI stores attribution metadata in git notes under its own Git AI standard. Selvedge is complementary: it adds explicitly recorded explanations, a query layer, and changesets, and can also emit the Agent Trace v0.1.0 record format.

Selvedge Git AI
Reasoning source Explanation supplied by an agent or person; timing depends on when it was recorded Attribution metadata
Granularity Entity — DB column, table, env var, dep, API route, function Line
Mechanism MCP server — an agent or person explicitly logs a change Agent-invoked checkpoint → Git notes at commit
Grouping Changesets — named feature/task slugs across many entities None
Prior attempts Yes — prior_attempts surfaces tried-and-rejected paths None
Storage SQLite Git notes
  • Reasoning, not just attribution. Git AI records who/what touched a line. Selvedge stores a supplied explanation of why alongside an explicitly logged change; it does not guarantee when that explanation was written.
  • A query layer. selvedge blame, diff, history, changeset, and search are first-class. Git-notes attribution has no equivalent CLI you can pipe.
  • Capture workflow. Git AI’s own README states it does not rely on git hooks: the agent calls git-ai checkpoint, and attribution lands in git notes at commit time. That is cooperative snapshotting — it records what the agent chose to check point, when it chose to. Selvedge also depends on explicit logging; the record may be written during or after the change.
  • Agent Trace interop. Selvedge also emits Agent Trace v0.1.0 records — see the Agent Trace interop page. (Git AI itself has since moved to its own Git AI standard rather than Agent Trace.)
  • Entity-level + changesets for querying the history of a column or a whole feature, not a line.

If all you need is line-level attribution metadata living in git notes under Git AI’s own standard — and you don’t need captured reasoning or a query layer — Git AI is a focused fit.