See what Coauthor reveals about your codebase

Coauthor runs locally against your own git repos. Here's what a typical analysis session looks like when you ask your AI assistant about your team.

Install Coauthor, add the MCP server, and ask these questions about your own repos. Every answer is grounded in your actual git history.

You ask:

"Who owns the authentication module?"

Coauthor responds:

src/auth/ -- Primary: sarah (67% structural ownership, 42 commits). Secondary: james (23%, mostly tests). Bus factor: 1.3 (concentrated).

Structural ownership goes beyond line counts. Sarah made the architectural decisions; James added test coverage. Both matter, but differently.

You ask:

"What were the highest-impact commits this sprint?"

Coauthor responds:

commit a3f8b2d (impact: 8.7/10) -- Refactored API gateway routing. Touched 3 files but affected 47 downstream consumers via import graph.

Impact scoring uses cross-file dependency analysis. A 3-file change that restructures an API scores higher than a 50-file rename.

You ask:

"Show me the bus factor for each subsystem"

Coauthor responds:

api/ 3.2 (healthy), auth/ 1.3 (risk), billing/ 1.0 (critical), docs/ 4.1 (distributed), tests/ 2.8 (adequate).

billing/ has a single contributor. If they leave, nobody else has touched that code. This is a concrete, actionable risk.

You ask:

"What's the team pattern breakdown?"

Coauthor responds:

2 specialists (deep auth, deep infra), 1 generalist (touches everything), 1 hub (connects frontend/backend), 3 peripheral (occasional fixes).

Pattern classification helps understand team dynamics. The hub contributor is often the most critical for cross-team coordination.

Try it on your own repos

Coauthor is free, open source (Apache-2.0), and runs entirely locally. Your git history never leaves your machine.