Documentation
Twenty-four documents, ordered by what you need first. The command reference is generated from the engine's own verb table and lives here.
Getting started
Designing your application on kevy
How to think in keys when you are used to thinking in tables.
The RDS→kevy modeling cookbook
Recipes that run: sessions, rate limits, queues, leaderboards, feeds.
Persistence
The append-only log, snapshots, and what survives a kill -9.
Tuning kevy
Shards, connections, and the two flags that actually matter.
Running it
Replication
One primary, N replicas, and the offset that tells you the truth.
Availability
Failover, epochs, and the writes we will not promise to keep.
Cluster
Why there is no cluster, and what to do instead.
--accept-shards N — fold sparse connections onto a subset of shards
The flag that bought 10.6% by giving fewer cores the listener.
Unix-domain socket (UDS) transport
Unix sockets: no TCP, no loopback, no kernel copy.
Async client
Pipelining, and why the engine has no async runtime.
Working with data
Secondary indexes (IDX. / idx_)
Secondary indexes: how they build, and what they cost.
Vector search (KIND ann)
KNN over your keyspace. No embedding model included.
Full-text search (KIND text)
BM25 full-text, and where it stops.
Views (VIEW. / view_)
Materialised views, kept fresh by the write path.
CDC — the change feed (FEED. / changes_since)
A change feed you can tail from another process.
Pub/sub
Channels, patterns, and what happens to a slow subscriber.
Lua scripting
EVAL, on an interpreter we wrote.