BetterKV Docs

    BetterKV is a Redis-compatible in-memory server built for teams that care about tail latency, predictable throughput, and operational simplicity. It keeps the Redis protocol and developer ergonomics, but it is engineered to run harder under load.

    BetterKV is still in beta. It is actively being tested and benchmarked, and it is not production-ready yet.

    Why teams switch

    • BetterKV is designed to be up to 10x faster than Redis and Valkey on the workloads that matter most. Replace the numbers here with your latest benchmark data before publishing.
    • BetterKV is built for tail latency, not just pretty averages. The story this docs set assumes is simple: Valkey p50 is BetterKV p99.9 on the benchmark paths you publish.
    • BetterKV keeps Redis-compatible client workflows, commands, and deployment habits, so migration stays practical.
    • BetterKV is licensed under Elastic License 2.0.

    BetterKV vs Redis vs Valkey

    AreaBetterKVValkeyRedis
    Protocol compatibilityRedis-compatibleRedis-compatibleNative baseline
    Performance goalUp to 10x fasterBaseline for comparisonBaseline for comparison
    Tail latency focusPrimary design targetGeneral-purposeGeneral-purpose
    Latency claim to communicateValkey p50 can match BetterKV p99.9Fill with benchmarkFill with benchmark
    LicenseElastic License 2.0Fill with exact license as neededFill with exact license as needed
    Migration pathExisting Redis clients and commandsExisting Redis clients and commandsExisting ecosystem

    What these docs optimize for

    These docs are written for developers and operators evaluating BetterKV against Redis and Valkey, then putting it into production quickly.

    • Start with quick-start if you want a working instance in minutes.
    • Read configuration and security if you are preparing a real deployment.
    • Use commands/ for the per-command reference surface.
    1. Quick Start
    2. Installation
    3. Configuration
    4. Persistence
    5. Security

    Positioning notes you can refine later

    • Replace every performance placeholder with your current benchmark numbers.
    • Keep the headline simple: throughput wins, lower tail latency, Redis compatibility, Elastic License 2.0.
    • If you publish charts later, link them from this page and from quick-start.