Back to Projects
AI·2026
RepoRAG
AI-powered internal tool that lets teams chat with their codebase using streaming responses to get repo health insights, limitations, and improvement suggestions.
Next.jsTypeScriptNode.jsPostgreSQLPineconeOpenAI APISSE
About This Project
RepoRAG is an internal organization tool that uses Retrieval-Augmented Generation to let developers have conversations with their code repositories. It indexes the entire codebase, analyzes structure and patterns, and provides health scores, identifies limitations, and suggests actionable improvements. Responses are streamed in real-time using Server-Sent Events (SSE) for a smooth, ChatGPT-like experience. Teams can ask natural language questions about their repo and get context-aware answers grounded in the actual code.
Key Highlights
- Real-time streaming responses using Server-Sent Events (SSE)
- Chat interface for natural language queries about the codebase
- Automated repo health scoring with detailed breakdowns
- Identifies code limitations and suggests improvements
- RAG pipeline with vector embeddings for accurate code retrieval
- Supports multiple repos within an organization