1
SiteShelf
SiteShelf is a personal web knowledge base that lets users save URLs, auto-extract and embed their content, and query the collected knowledge using natural language. It uses semantic retrieval with vector search and re-ranking to surface the most relevant context quickly.
Key Features
- Built an AI-powered semantic retrieval system for web content where users save URLs and query extracted knowledge using natural language.
- Designed an end-to-end pipeline for web scraping, content extraction, embedding generation, and vector indexing to enable fast semantic search.
- Implemented vector similarity search with re-ranking to improve accuracy and contextual relevance of results.
- Developed a scalable edge architecture using Cloudflare Workers to support fast ingestion and low-latency queries at scale.
- Built a clean, responsive UI for saving links, managing a library, and searching across stored content.
Tech Stack
HonoCloudflare WorkerSupabaseVector DBPostgresOpenAI