What you get * Clone + parse + index any number of repos (20+ languages supported) * Generate embeddings, store them in a vector database, enable semantic search by intent (not just keyword matching) * Ask natural language questions like “how does authentication work?” — get relevant file snippets returned
Why it matters * Navigating large or polyrepo codebases manually is slow and error-prone * Semantic search helps you find relevant code even if you don’t remember exact file names or code paths * REST API + docker-compose deployment lets you self-host quickly and integrate it with existing workflows
If you work with large or multiple codebases, rephole can save you time and make code navigation easier. Feedback, issues or PRs welcome