# Hanzo KV — Hanzo AI

> Managed Redis/Valkey for caching, sessions, rate limiting, and pub/sub.

In-memory data store

# Hanzo KV

In-memory key-value store

Values live in memory, so a read is a network hop and a hash lookup — nothing waits on a disk. KV speaks RESP2 and RESP3 on port 6379, so the client library you already have connects to it. Reach for it for caches, sessions, rate limits, job queues and leaderboards.

RESP

2 and 3

Sentinel

Failover

AOF

Persistence

Cluster

Sharding

[Get Started](https://docs.hanzo.ai/docs/kv)[GitHub](https://github.com/hanzoai/kv)

## Pick a structure, not a schema

A leaderboard is a sorted set. A rate limit is a counter with an expiry. A queue is a list.

### Nothing touches a disk

Reads and writes are answered out of memory, so the time you spend is the network and a hash lookup. That is the whole reason to put a cache in front of a database.

### The structures do the work

Strings, hashes, lists, sets, sorted sets, streams, bitmaps, HyperLogLog and geospatial indexes — each with the commands that make sense for it. Lua scripts run several commands as one atomic step when a single command is not enough.

### Pub/Sub

Publish to a channel and every subscriber has it, including subscribers matching a pattern like orders.*. Keyspace notifications turn a key expiring into a message something else can act on.

### Keys that expire

Put a TTL on any key and it goes away by itself. When memory fills, you choose what leaves first — least recently used, least frequently used, shortest time to live, or only keys that already carry an expiry.

### Two ways to survive a restart

An RDB snapshot writes the whole dataset on an interval; the append-only file records every write as it happens. Run either or both, and check a file before you trust it with kv-check-rdb and kv-check-aof.

### ACLs on commands and keys

A user is granted specific commands and specific key patterns, so the service that only reads cache keys cannot reach FLUSHALL. TLS on the wire.

## Any RESP client connects

cache.ts

```
import Redis from "ioredis" const kv = new Redis("redis://kv.hanzo.ai:6379") // Caching await kv.set("user:123", JSON.stringify(user), "EX", 3600) const cached = await kv.get("user:123") // Rate limiting const count = await kv.incr("ratelimit:api:usr_123") await kv.expire("ratelimit:api:usr_123", 60) // Pub/Sub kv.subscribe("events") kv.on("message", (channel, message) => { console.log(channel, message) })
```

Open Source Revenue Sharing

### Up to 5% of compute goes back to open source

Every deployment is SBOM-verified. Contributors to Valkey earn a share of compute revenue — transparent, on-chain, and customizable by the community.

[Connect & Earn](https://hanzo.ai/open-source/dividends)[Learn More](https://hanzo.ai/open-source)

Open Source Compute Dividends

### Up to 5% of compute goes back to OSS authors

Every Hanzo deployment tracks software dependencies via SBOM. When your code powers compute on Hanzo, the authors get paid — automatically.

USDHanzo NetworkETHSOLUSDCmore

[Connect GitHub](https://console.hanzo.ai?ref=oss-dividends)[How payouts work](https://hanzo.ai/open-source/dividends)

## Put it in front of the database

Free tier includes 256 MB. Provision in seconds.

[Get Started](https://docs.hanzo.ai/docs/kv)[View on GitHub](https://github.com/hanzoai/kv)

## Open source

License: BSD-3-Clause[hanzoai/kv](https://github.com/hanzoai/kv)

## Get KV

Distributed key-value store

[Deploy to Cloud](https://console.hanzo.ai/deploy)[Self-host](https://docs.hanzo.ai/docs/kv)