Your data.
At the speed
of thought.
Less waiting. More building.
A Redis-compatible in-memory data store written in Go. Your favorite clients, a powerful engine, and everything you need in one binary.
curl -fsSL …/install.sh | bashConcurrent access. Data stays in the fast lane.
ONE SMALL BINARY.
A WHOLE LOT INSIDE.
FAST IS JUST THE BEGINNING
A simpler stack.
A stronger foundation.
From your first cache key to your busiest day.
An engine that gives you room to grow.
A fast lane
for every request.
256 shards distribute the work. Goroutines handle connections. Your data lives in memory, right where speed happens.
Look under the hoodKeep your clients.
Meet your new engine.
Connect over RESP with the Redis libraries you already know. Familiar commands. A fresh foundation.
Good boundaries.
Built in.
TLS encryption, authentication and access control lists. Give each application the access it needs.
TLS + per-user ACLMemory for speed.
Persistence for peace of mind.
Keep it ephemeral or turn on AOF and snapshots. Choose the durability your workload needs.
Your data. Your decision.One binary.
All the essentials.
HTTP API, Prometheus metrics, Lua scripting, transactions and a plugin architecture. Ready when you are.
Explore every capabilityFROM ZERO TO YOUR FIRST PONG
Ready when
your idea is.
Install it. Start it. Connect.
Keep the ceremony to a minimum.
Your machine.
Your way to run.
A native binary for your platform or a container in your stack. Pick your starting point.
curl -fsSL https://raw.githubusercontent.com/CacheStorm/CacheStorm/main/scripts/install.sh | bash$ redis-cli -p 6380
127.0.0.1:6380> PING
PONG
127.0.0.1:6380> SET hello "world" EX 3600
OK
127.0.0.1:6380> GET hello
"world"Your existing tools. A familiar first connection.
MORE THAN KEY. VALUE.
One engine.
So many possibilities.
Sessions. Leaderboards. Live events.
Choose the shape your data needs.
A home for the hot path.
Fast key-value access with optional expiry. Keep sessions, cached responses and counters close.
user_128TTL 3600sSET session:42 "user_128" EX 3600
GET session:42
TTL session:42OK
"user_128"
(integer) 3600Examples show expected command shapes and illustrative output. No server runs in your browser.
MADE FOR THE HOT PATH
Small footprint.
Storm-sized potential.
Purpose-built storage. Carefully measured performance.
Sequential · in-memory store
Parallel · 32 logical processors
End-to-end · TCP round trip
Repository-reported benchmarks · AMD Ryzen 9 9950X3D · 64 GB RAM · Windows 11.
Storage benchmarks exclude networking. TCP results include the full round trip. Performance varies by workload.
WHAT WILL YOU BUILD?
Your next big thing.
A little less infrastructure.
CacheStorm is an open-source, Redis-compatible in-memory data store written in Go. It combines a 256-shard storage engine with rich data structures, an HTTP API, observability and optional persistence in a single binary.
Yes. CacheStorm speaks the RESP protocol and supports familiar Redis commands. Point your client at port 6380, then test your application’s command usage and behavior before migrating a production workload.
The numbers are reported in the upstream repository. The fastest results measure direct in-memory store operations without networking. End-to-end TCP results are listed separately on the benchmarks page. Results depend on hardware and workload.
You can enable append-only files (AOF) and snapshots. The default example configuration has persistence disabled, so enable it explicitly and test backup and recovery for your deployment.
Yes. The CacheStorm source is licensed under Apache 2.0. You can use, modify and distribute it commercially subject to the license terms, including the required notices.
Start with the quick start, connect using redis-cli, then try the commands your application needs. The configuration, security and monitoring guides cover the next steps for running your own instance.
OPEN SOURCE. OPEN POSSIBILITIES.
Great ideas
deserve a fast start.
Pure Go. Apache 2.0. Yours to build on.