Overlay Compatible Integrations

Works with your
existing cache.

Cachee doesn't replace your cache provider. It sits in front as an L1 acceleration layer — delivering ~10us cache hits while your existing backend handles durability. Zero code changes. Zero migration.

Start Free Overlay How It Works

How It Works

Three steps. Ten minutes.

No migration, no data movement, no code changes. Cachee sits between your app and your existing cache.

Deploy Cachee Proxy

One Docker container. Set your API key and point UPSTREAM at your existing cache endpoint. Takes 3 minutes.

Swap One Connection String

Change your app's REDIS_HOST from your current provider to the Cachee proxy. Your existing Redis client works as-is.

L1 Serves Hot Data in ~10us

Cachee's Tiny-Cachee engine caches your hottest keys in-memory. Misses forward transparently to your existing backend.

Compatible Providers

Overlay any of these. Instantly.

Cachee speaks Redis protocol natively. Any RESP-compatible endpoint works out of the box. HTTP-based providers like CloudFlare KV use our built-in protocol adapter.

Most Popular

AWS ElastiCache

Redis • RESP Protocol

The most common Cachee overlay target. ElastiCache delivers <1ms in same-VPC, but Cachee's L1 serves hot keys 100x faster at ~10us with zero extra infrastructure.

<1ms

ElastiCache alone

~10us

With Cachee L1

90%+

Cost reduction

# docker-compose.yml UPSTREAM=redis://my-cluster.abc.cache.amazonaws.com:6379 CACHEE_API_KEY=${CACHEE_API_KEY}

Edge + L1

CloudFlare Workers KV

HTTP KV • Protocol Adapter

CloudFlare KV is great for global edge reads (10-50ms), but Cachee's L1 brings hot-key latency down to ~10us. Reduces CF API calls by 90%+, cutting both latency and cost.

10-50ms

CF KV alone

~10us

With Cachee L1

1000x

Faster hot keys

# CloudFlare KV via protocol adapter UPSTREAM=cloudflare://ACCOUNT_ID/NAMESPACE_ID UPSTREAM_CF_TOKEN=${CF_API_TOKEN}

Redis Cloud

Redis Labs • RESP Protocol

Redis Cloud provides managed Redis with active-active geo-distribution. Cachee adds an L1 layer that eliminates network round-trips for your hottest keys.

1-5ms

Redis Cloud alone

~10us

With Cachee L1

75%

Fewer API calls

# Redis Cloud endpoint UPSTREAM=redis://redis-12345.c1.us-east-1.ec2.redislabs.com:12345 CACHEE_API_KEY=${CACHEE_API_KEY}

Azure Cache for Redis

Microsoft Azure • RESP Protocol (TLS)

Azure Cache uses TLS by default on port 6380. Cachee connects via rediss:// and adds L1 acceleration without changing your Azure infrastructure.

<2ms

Azure alone

~10us

With Cachee L1

TLS

Native support

# Azure Cache (TLS on port 6380) UPSTREAM=rediss://my-cache.redis.cache.windows.net:6380 CACHEE_API_KEY=${CACHEE_API_KEY}

GCP Memorystore

Google Cloud • RESP Protocol

Memorystore provides fully managed Redis on Google Cloud. Cachee overlays it with an L1 layer, reducing read latency and Memorystore instance costs.

<1ms

Memorystore alone

~10us

With Cachee L1

80%

Fewer reads

# GCP Memorystore (private IP) UPSTREAM=redis://10.128.0.3:6379 CACHEE_API_KEY=${CACHEE_API_KEY}

Upstash

Serverless Redis • RESP Protocol (TLS)

Upstash provides serverless Redis with per-request pricing. Cachee's L1 cuts Upstash API calls by 90%+, dramatically reducing your per-request costs.

5-15ms

Upstash alone

~10us

With Cachee L1

90%

Cost savings

# Upstash (TLS) UPSTREAM=rediss://usw1-example.upstash.io:6379 UPSTREAM_PASSWORD=${UPSTASH_TOKEN}

Any Redis Endpoint

Self-Hosted • RESP Protocol

Running your own Redis? Docker Redis? DigitalOcean, Heroku, Render, or bare metal? If it speaks Redis protocol, Cachee can overlay it.

Any

RESP endpoint

~10us

L1 hit latency

Code changes

# Any Redis-compatible endpoint UPSTREAM=redis://your-redis-host:6379 CACHEE_API_KEY=${CACHEE_API_KEY}

Before & After

What changes when you add Cachee.

Your existing cache stays exactly where it is. Cachee adds an L1 layer in front.

Metric	Your Provider Alone	Your Provider + Cachee
Hot key read latency	1-50ms (network dependent)	~10us (100-5000x faster)
Cache hit rate	70-85% typical	95%+ (AI pre-warming)
Provider API calls	100% of reads hit backend	90% served from L1
Monthly provider cost	Full price	60-90% reduction
Code changes required	N/A	Zero — swap connection string
Data migration	N/A	None — overlay, not replace
Rollback plan	N/A	Instant — revert connection string
Setup time	N/A	10 minutes

Your cache. Supercharged.

Pick your provider in the setup wizard. Cachee handles the rest.

Start Free Overlay See All Deploy Options

Works with yourexisting cache.

Three steps. Ten minutes.

Overlay any of these. Instantly.

What changes when you add Cachee.

Your cache. Supercharged.

Works with your
existing cache.