Hey DEV community! 👋

I’ve been working on a side project that turned into something surprisingly useful, fun, and very Rust-y.

Meet MemCloud — a distributed in-memory data store that lets multiple machines on your LAN pool their RAM into a single ephemeral “memory cloud.”

Use case: Have a Mac, a Linux machine, and a spare mini-PC sitting around?

MemCloud turns them into one big RAM cache — automatically.

---

💡 Why I Built This

I often run ML experiments, dev servers, and log processors that overflow RAM on one machine while another machine sits idle right next to it.

I wanted a tool that:

• works offline
• runs locally
• requires zero configuration
• discovers peers automatically
• lets me store/load data across devices in milliseconds

So I built MemCloud, a tiny Rust daemon + CLI + SDKs that create a peer-to-peer RAM mesh on your LAN.

---

⚡️ Key Features

🕸️ P2P RAM Pooling

Every memnode contributes its RAM to the cluster.

A write on Machine A can be read from Machine B in under 10ms.

🔍 Zero-Config Discovery (mDNS)

Just start the daemon — peers auto-discover each other.

No IPs, no ports, no YAML files, no Kubernetes.

🌐 Works Fully Offline

No cloud. No accounts. No central server.

⛓️ CLI + SDKs

• memcli for terminal workflows
• Rust SDK for systems work
• TypeScript SDK for JS/Node devs

🔑 Two Storage Modes

• Block Store: raw bytes & streams
• Key-Value Store: Redis-style set / get

---

🧱 Architecture

Each node runs a small daemon (memnode).

SDKs and CLI talk only to the local daemon.

The daemon handles routing and storage across peers.

---

⚙️ Installation

Quick Install (macOS & Linux)

curl -fsSL https://raw.githubusercontent.com/vibhanshu2001/memcloud/main/install.sh | sh

🛠️ Build from Source

git clone https://github.com/vibhanshu2001/memcloud.git
cd memcloud
cargo build --release

👨‍💻 Open Source

---

I’d love feedback on:

• performance ideas
• networking improvements
• memory/eviction strategies
• real-world use cases

Thanks for reading!

— Vibhanshu