// the take
nStore is a tiny embedded key/value store for Node.js that appends writes to a single file and keeps an in-memory index of byte offsets for fast reads. It's aimed at small scripts or prototypes that want persistence without standing up a real database — think session storage or a quick local cache, not anything serving real traffic.
The append-only format is a genuinely nice design for this scale: writes are simple sequential appends, the file is always in a consistent state, and you can back it up with rsync while it's live. The compaction filter hook (pruning stale docs like expired sessions during compaction instead of issuing individual removes) is a clever touch that avoids bloating the file with delete markers. The query layer is a sensible optional add-on rather than baked into the core, keeping the base store minimal.
Dead since 2017 — nine years with no commits, and the README itself says 'this library is still under development, there are bugs, APIs will change,' which was never resolved. There are no real indexes; every query is a full in-memory scan over documents reconstructed from the index, so it won't hold up past a small dataset. The entire key index lives in memory, so RAM scales linearly with document count regardless of how much you compact the file on disk. It's strictly single-process with no concurrency story (locking, multi-process access, or crash recovery beyond 'the file is consistent'), so it's a non-starter for anything that needs more than one writer.
We read the README, the directory tree and a key source file in October 2026. Nobody pays to be listed. How we judge →