I built it out of necessity: I needed a fast, in-process database to handle large-scale entity resolution, full-text filtering, and relationship traversal. All without an external DB server. Nothing existing fit that niche, so LightDB was born.

What it offers:

Key-value access (RocksDB, LMDB, MapDB, etc.)

A concise Scala SQL-like DSL

A built-in graph traversal engine

Lucene-powered full-text search & faceting

Runs fully embedded in your app. No servers, no sockets

Blazingly fast: processing jobs that used to take days now finish in hours

Everything is composable, strongly typed, and designed for performance-critical JVM applications.

I’d love to hear feedback, especially if you’ve ever tried to use something like RocksDB, SQLite, or Lucene in a Scala app and found the integration painful.