BjornMelin/stardex

Uses a monorepo architecture with a Next.js frontend and FastAPI backend—the backend vectorizes repository descriptions via TF-IDF and applies K-means, hierarchical clustering, or PCA-based clustering via scikit-learn, with tunable parameters exposed through the UI. Fetches starred repos directly from GitHub's REST API in the browser, then sends metadata to the clustering service for semantic grouping, supporting multi-user star aggregation and fine-grained filtering by language, topics, and star count.