RepoGraph enables LLMs to understand entire code repositories instead of isolated functions, bridging the gap for complex software engineering tasks.

• Repository-level representation: Transforms entire codebases into navigable knowledge graphs
• Contextual awareness: Helps AI understand dependencies, relationships, and architecture across multiple files
• Enhanced capabilities: Significantly improves performance on repository-level tasks including debugging, refactoring, and feature integration
• Practical application: Addresses real-world software engineering scenarios that require holistic codebase understanding

This research advances AI software engineering by equipping LLMs with the repository-wide context needed for modern development workflows, making AI more effective for professional software engineering teams.

RepoGraph: Enhancing AI Software Engineering with Repository-level Code Graph