Welcome to the Valkey Community Docs
A practitioner-focused guide to Valkey, the BSD-licensed Redis 7.2.4 fork stewarded by the Linux Foundation.
What Valkey is
Valkey is an in-memory key-value data store: strings, hashes, lists, sets, sorted sets, streams, bitmaps, HyperLogLogs, geospatial indexes, and (via modules) JSON and vectors. If you have ever used Redis, you already know Valkey — the wire protocol (RESP2/RESP3) and the client libraries are the same.
Valkey was forked from Redis 7.2.4 in March 2024 after Redis Inc. switched the project to SSPL/RSAL. A group of long-time maintainers from AWS, Google, Oracle, Ericsson and Snap created Valkey under the Linux Foundation within eight days. Today Valkey is the engine behind AWS ElastiCache, Google Memorystore, Aiven, and many self-hosted deployments worldwide.
Latest release: Valkey 9.1 (May 2026), LTS-track. The 7.2 line is still maintained for conservative deployments.
Pick a path
5-minute quickstart
Run Valkey with Docker or a package manager, ping it, store your first key.
For beginners
What an in-memory KV store actually is, when to use one, and what Valkey gives you over a database.
For AI developers
Vector search, semantic cache, and agent memory in one process. Valkey as the AI data layer.
On the cloud
Provider matrix: AWS, GCP, Aiven, Oracle, DigitalOcean, and what each one actually gives you.
Versions & roadmap
7.2 / 8.0 / 8.1 / 9.0 / 9.1 — what changed, what's coming, which to pick.
Operations
Install, configure, persist, replicate, cluster, secure, observe — in one place.
Why this site
The official Valkey documentation at valkey.io is excellent but reference-shaped. This site is community-written, problem-shaped, and Chinese-first — but with an English mirror because the surrounding ecosystem (and many error messages you will Google) is in English.
If you find a mistake, open an issue on GitHub.