惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

J
Java Code Geeks
GbyAI
GbyAI
Google DeepMind News
Google DeepMind News
Jina AI
Jina AI
B
Blog
aimingoo的专栏
aimingoo的专栏
酷 壳 – CoolShell
酷 壳 – CoolShell
T
The Blog of Author Tim Ferriss
Last Week in AI
Last Week in AI
月光博客
月光博客
H
Help Net Security
V
Visual Studio Blog
量子位
A
About on SuperTechFans
博客园 - Franky
人人都是产品经理
人人都是产品经理
N
Netflix TechBlog - Medium
云风的 BLOG
云风的 BLOG
雷峰网
雷峰网
Martin Fowler
Martin Fowler
Microsoft Security Blog
Microsoft Security Blog
博客园 - 叶小钗
P
Proofpoint News Feed
MongoDB | Blog
MongoDB | Blog

Hacker News: Show HN

PurrrrrFocus: Pomodoro Timer App - App Store Workflow Engine — Multi-Step Orchestration for Bun RapidPhoto: Pro Photo Editor App - App Store GitHub - DheerG/swarms: Achieve extraordinary results with claude code across a variety of tasks SPICE simulation → oscilloscope → verification with Claude Code — Lucas Gerads Show HN: VCoding – A 5 MB native Windows IDE with no dynamic dependencies Show HN: LLMs don't hallucinate because they're bad at math, it's the format GitHub - Agent-FM/agentfm-core: AgentFM is a peer-to-peer network that turns everyday computers into a decentralized AI supercomputer. AgentFM lets you run massive AI workloads directly across a global mesh of idle CPUs and GPUs. Show HN: Tracking Top US Science Olympiad Alumni over Last 25 Years GitHub - Potarix/agent-hub: One place to talk to all your agents Show HN: Runtime security for AI agents(injection,tool abuse, data exfiltration) GitHub - dubeyKartikay/lazyspotify: Terminal Spotify client for macOS and Linux GitHub - the-banana-tool/king-louie: Easy to use GUI Personal AI Assistant. Win/Linux/Mac. Show HN I made my vacation rental bookable by AI agents–no Airbnb, 0% commission GitHub - basteez/jsf-autoreload: maven plugin to enable hot reload on jsf projects uvm32/hosts/host-gdbstub at main · ringtailsoftware/uvm32 GitHub - labsai/EDDI: Config-driven engine that turns JSON into production-grade AI agents. Multi-agent orchestration, 12+ LLM providers, MCP/A2A protocols, RAG, persistent memory, and enterprise compliance (EU AI Act, GDPR, HIPAA). Built on Quarkus. GitHub - glitchnsec/fortyone-oss: AI Executive Assistant Platform Quickstart | Alien GitHub - muxshed/shed: One stream in, or many. Every destination, simultaneously. No cloud middleman, no per-channel fees, no limits. GitHub - ocrbase-hq/ocrbase: 📄 PDF/IMG ->.MD/JSON Document OCR API for PaddleOCR and GLMOCR. Self-hostable. GitHub - impactjo/home-memory: MCP server that lets your AI assistant remember everything about your home. GitHub - Sets88/dbcls: DbCls is a powerful terminal database client that supports various databases GitHub - neptun2000/heor-agent-mcp GitHub - SeanFDZ/macmind: Single-layer transformer in HyperTalk for the classic Macintosh RollQuation: Math Puzzles - Apps on Google Play GitHub - dropbox/witchcraft Show HN: Agent-cache – Multi-tier LLM/tool/session caching for Valkey and Redis GitHub - opentalon/opentalon: OpenTalon is an open-source platform built from the ground up in Go as a robust alternative to OpenClaw LinkedIn™ 职位抓取工具 - Chrome 应用商店
GitHub - GianIac/numax: The beginning of a new runtime model
gianiac · 2026-06-17 · via Hacker News: Show HN

NUMAX

Docs Whitepaper Roadmap

A portable runtime for distributed apps. Written in Rust.

Three things, and only three:

  1. Runs WebAssembly modules in an isolated sandbox.
  2. Has a local embedded key/value datastore, state lives next to the code.
  3. Syncs state across nodes with CRDTs and gossip.

You write a WASM module. numax runs it. The state is there. Sync just happens.

Status: v0.1.0 - first stable Numax release line for controlled, non-critical workloads. It works, it's tested, and the remaining limits are documented. See the Roadmap.


Why

Building distributed software today is heavier than the problem it's trying to solve. Containers, orchestrators, remote databases, ad-hoc sync layers, three different toolchains depending on where the code runs.

numax tries a different path: keep the runtime tiny, keep the state local, let CRDTs handle convergence. The hard parts of distributed systems don't disappear, but you stop paying for the ones you didn't actually need.


Quickstart

Quickstart in 5 Minutes.

For now, build from source:

git clone https://github.com/GianIac/numax
cd numax
cargo build --release

Run the distributed_counter example on two nodes:

# Node A
nx run distributed_counter.wasm \
    --listen 0.0.0.0:9000 \
    --datastore-path ./data-a -v

# Node B
nx run distributed_counter.wasm \
    --listen 0.0.0.0:9001 \
    --peer 127.0.0.1:9000 \
    --datastore-path ./data-b -v

The two nodes find each other, sync, and converge. You don't have to do anything else.

You can also move the node settings into TOML files and keep the command line focused on run-specific options:

# node-a.toml
[storage]
datastore_path = "./data-a"

[network]
listen = "0.0.0.0:9000"
peers = ["127.0.0.1:9001"]
serialization_format = "bincode"

[discovery]
mode = "static"
nx config validate --config node-a.toml
nx config show --config node-a.toml --effective

nx run distributed_counter.wasm \
    --config node-a.toml \
    --settle-for 5s \
    --print-gcounter counter:visits

Runtime precedence is explicit: CLI flags override NX_* environment variables, environment variables override the TOML file, and the file overrides runtime defaults. The distributed examples below include full two-node TOML setups.


Writing a guest module

Your first module

A minimal module using the local datastore:

use nx_sdk::{db, log};

#[no_mangle]
pub extern "C" fn run() {
    db::set("hello", b"numax").unwrap();
    log("done.");
}

Or with a replicated CRDT counter:

use nx_sdk::{log, crdt::gcounter};

#[no_mangle]
pub extern "C" fn run() {
    gcounter::inc("visits", 1).unwrap();
    let v = gcounter::value("visits").unwrap();
    log(&format!("visits: {}", v));
}

Same module, any node. State stays local. Sync happens through the runtime.


Learn more


A small ask

If numax interests you - if you think the idea is worth something - drop a star !

Right now it's pretty much the only signal I have to understand whether this is worth pushing further.


Try it. Break it. Tell me.

numax is in its first stable release line. It is usable, but still early: focused feedback matters a lot.

  • Clone it, run the examples, see if the two nodes really converge on your machine.
  • Write a tiny module of your own and try to break the sandbox or the sync.
  • If something behaves in a way you didn't expect - open an issue. Even a small one. Especially a small one.
  • If you have an opinion on the design, the host API, the CRDT model - open an issue for that too.

There's no community to pretend already exists. There's a project, an idea, and a door that's open. If you walk through it now, you're early. That's the best moment to leave a mark.

ps: If you'd like to help, take a look at CONTRIBUTING.md.

  • GianIac

License: Apache 2.0