Understanding how your code changes affect other parts of your codebase is crucial during reviews, but manually tracking these relationships is time-consuming and error-prone. Code graph analysis automatically maps connections between files to surface relevant context exactly when you need it.

Definition context

Shows exact files containing the User interface when you modify a function that returns User[], eliminating guesswork about type definitions.

Reference tracking

Identifies the 8 components importing Button.tsx when you change its props, preventing breaking changes from reaching production.

Change patterns

Learns from your team’s commit history that api/users.ts and types/user.ts frequently change together, surfacing hidden dependencies.

​
How it works

Code graph analysis performs static analysis on your entire codebase during each review, parsing abstract syntax trees to understand symbol definitions and import relationships. Unlike basic grep searches, it understands language semantics - when you modify a TypeScript interface, it knows which files import that interface and how it’s used. The system learns from your team’s development patterns, analyzing commit history to identify files that frequently change together even when there’s no obvious code relationship.
1

Review any pull request

Code graph analysis activates automatically - no configuration needed.
2

Check the review details

Look for “Additional context used → Code definitions” in the review status to see graph analysis in action.
3

See contextual insights

Review comments will reference related files and highlight cross-file implications of your changes.
Available on Lite tier and above. The feature also powers Agent Chat for codebase-wide questions and analysis.