Hey everyone! After months of tinkering, I’m excited to share readwise-vector-db—an open source project that transforms your Readwise highlights into a blazing-fast, self-hosted semantic search engine.

Why? I wanted a way to instantly search my entire reading history—books, articles, PDFs, everything—using natural language, not just keywords. Now, with nightly syncs, vector search API, Prometheus metrics, and a streaming MCP server for LLM clients, it’s possible.

Key features:• Full-text, semantic search of your Readwise library (local, private, fast)• Nightly sync with Readwise—no manual exports• REST API for easy integration with your tools and workflows• Prometheus metrics for monitoring• Streaming MCP server for LLM-powered apps

It’s Python-based, open source (MIT), and easy to run with Docker or locally. If you want to own your reading data, build custom workflows, or experiment with local LLMs, give it a try.

Would love feedback, questions, and ideas for next steps!