nalgeon/redka

Redka is a project that recreates the core features of Redis using a SQL database as its storage engine. Redis is a popular system for storing data in memory, often used for caching and quick lookups. Redka offers the same programming interface as Redis, so code written for Redis will largely work with Redka without changes, but the data lives in SQLite or PostgreSQL instead of RAM. The practical difference is that Redka data does not have to fit in memory. Because it uses a real database on disk, you can store more than your available RAM would normally allow. You also get database-style guarantees: changes are either fully saved or fully rolled back, which makes it safer for certain kinds of operations where partial updates would cause problems. Redka supports the five core data types that Redis users typically rely on: plain text values, ordered lists, unique sets, field-value maps, and scored sorted sets. It also handles key expiration and transactions, so the behavior matches what Redis users expect from those commands. You can run Redka in two ways. One option is as a standalone server that behaves just like a Redis server, so any Redis client library in any programming language can connect to it. The other option is to embed it directly inside a Go application as a library, which means no separate server process is needed at all. The embedded mode is particularly handy for testing, since each test can spin up its own isolated in-memory database without needing a real Redis installation. The project is described as stable but in maintenance mode, meaning the author is not adding new features. It is best suited for development environments, lightweight production use cases, or situations where a team already uses SQLite or PostgreSQL and wants Redis-like data structures without managing a separate Redis server.